Tiruppur Kumaran College for Women - PG & Research Department of Computer Science
Julie M. David
This paper focuses on detecting brain tumor in human beings. A brain imaging technique called Magnetic
Resonance Imaging (MRI) is used to collect the images of the brain. MRI scan can provide information about the blood circulation throughout the body and blood vessels and also enabling the detection of problems related to the blood circulation in the brain. In this work, first the image is segmented using K-means clustering algorithm. Next performed a feature extraction on the segmented image using Discrete Wavelet Transform (DWT). Finally, the extracted features are classified using Support Vector Machine (SVM). In SVM I first create a classification model and then classify the new data based on the trained dataset. The classifier created by the proposed research work produced an accuracy of 90% and is efficient.
V.Indhumathi, Dr.G.M.Nasira
In the current situations, complex real time applications depend on decentralized location for handling many practical problems. Therefore distributed clusters are used with unlimited computing resources. Different computers are interconnected through high-speed networks to provide efficient high performance computing platform for solving complex real time applications such as Heterogeneous clusters. In such environment, the constraints of distributed systems can be switched to develop the underutilized computing resources in all regions throughout the world for dispersed jobs. Even though Resource Scheduling, Load balancing and fault tolerance is well known research field, there are lots of new challenges in grid environments such as receiving up normal arrival of jobs, clustering jobs to resources, tolerate fault during load balancing are the main consideration of our proposed Benders decomposition.
C.Sandhiya, M.Sasikala
Cloud information stockpiling rethinks the issues focused on client's out-sourced (information that is not put away/recovered from the costumers claim servers).In this work we watched that, from a client's perspective, depending upon a performance SP for his outsourced information is not extremely encouraging. To address these issues in this paper, we proposed the systems for circulation of information among the accessible SP s in the market, to furnish clients with information accessibility and also unwavering quality. Our proposed approach will give the distributed computing clients a choice model, which gives a superior unwavering quality and accessibility by circulating the information over different cloud specialist organizations such that, none of the SP can effectively recover and utilize it.
R.Vidhu, S.Kiruthika, C.B.Lakshmi
Oral Squamous Cell Carcinoma (OSCC) constitutes the 8th most common neoplasm in humans. OSCC results from a combination of risk habit factors such as tobacco use, betel-quid chewing, alcohol consumption and genetic damage that leads to DNA alterations in key cellular genes. Diagnosis of oral cancer at its early stage will reduce the mortality rate. The diagnosis requires data collection from patients. Data mining techniques, such as pattern association, classification and clustering, are now frequently applied in cancer and gene expressions correlation studies. In this study, a framework for learning the structure of oral cancer genetic network based on DBNs is proposed. This approach improves the prediction level of the oral cancer and is tested with the gene data is set to prove the increase of classification accuracy and reduce execution time compared to the existing technique.
N.Shanmugapriya
Medical image processing is one of the most eminent image processing fields in this era. This is because of the big revolution in information technology that is used to diagnose many illnesses and saves patients’ lives. There are many image processing techniques used in this field, such as image reconstructing, image compression, segmentation and many more. Image compression is a mandatory step in many image processing procedures. Image compression is a most significant tool which reduces the burden of storage and transmission over network. Hence, the medical images need to be transmitted very fast and it required to store with a minimum capacity. Thus, image compression is used to reduce the redundancies and irrelevant information in image and represents it in shorter manner to achieve efficient archiving and transmission of images. Image compression is the process of reducing irrelevant and the redundancy of the image data in order to store or transmit data in an effective manner. Image compression minimizes the size of an image (in bytes) without degrading the quality of an image to an acceptable level. In this paper, we have presented the work done in the field of medical image compression.
B.Ramya, S.Hemalatha, M.S.Kokila
In this paper, the mining of various dimensional association rules with rough set approach is investigated as the algorithms RSSAR,RSMAR,RSHAR.The RSSAR single dimensional association rule algorithm. It is used to scan the transactional databases or dataset many time to find frequent itemsets. It is proposed for mining hybrid dimensional association rule using multiindex structures for storing multidimensional , Interdimensional and Intra dimensional frequent item set and it stores all frequent 1 – item sets after scanning the entire database first time in the temporary table for compression of the transactional datasets. In The RSMAR algorithm is constituted of two steps mainly. At first, to join the participant tables into a general table to generate the rules which is expressing the relationship between two or more domains that belong to several different tables in a database. Then we apply the mapping code on selected dimension, which can be added directly into the information system as one certain attribute. To find the association rules, frequent item sets are generated in second step where candidate itemsets are generated through equivalence classes and also transforming the mapping code in to real dimensions. The searching method for candidate item set is similar to apriori algorithm. The analysis of the performance of algorithm has been carried out.
S.Niranjana, J.Shymali, M.Narmatha
Famously known as the world’s first virtual brain, “BLUE BRAIN” is a very appropriate application of an artificial intelligence human brain. That means a machine can function as human brain. It’s a known fact that human does not live for a decade but the information contained in his mind could certainly be stored for a decade with this technology. So with advancement in technology, even after a person is dead, the virtual brain will serve as the man. Therefore, the main idea behind this is uploading human brain into machine. Hence this research paper consists of the concepts of Blue Brain, its requirements, why and how it works, applications, and experiments.
T.Kavipriya, N.Geetha
Data mining is an activity of extracting some useful knowledge from a large data base, by using any of its techniques. Due to a rapid advancement in the electronic commerce technology, the use of credit cards has increased. As credit card becomes the most popular mode of payment for both online as well as regular purchase, cases of credit card fraud also rising. Data mining techniques could be used to detect the credit card fraud detection. The main goal of this paper is compare the data mining techniques, such as Simple K-means, Hidden Markov Model, Bayesian Network, KNN algorithm and Outlier detection.
P.Chitra, Dr.T.R.Ranganayaki
Mobile wireless sensor networks are the new generation of network and the networks are organized by the form of Double -layer which can adapt to the application of the sink node and the sensor node with mobility. The main objectives of this method are uniform distribution, increase of network life time, decrease of traffic load and load balance and security in data transfer in reliability of formed path. This starts from network architecture and is divided into high-end node layer, DEN (Data Encryption node) layer and low-end node layer. The high-end node are responsible for the data routing, DEN is used for secure the data and the low- end node is used for sensing and reporting data. This approach will be for data transfer in security and reduce the energy consumption.
S.Sandhya, Dr.G.M.Nasira
Weather forecasting is the application of science and technology to predict the state of the atmosphere for a given location. Prediction is a challenging task and that too for weather is even more complex, because it depends on various parameters to predict the dependent variables like temperature, rainfall, humidity, wind speed and direction, which are changing from time to time and weather calculation varies with the geographical location along with its atmospheric variables. In this paper, we analyse the data mining techniques in forecasting weather.
V.Hemalatha, C.Usha nandhini
Classification of coronary heart disease can be valuable for the medical practitioners in the event that is automated with the end goal of quick finding and exact result. Recently, researchers have used different classification and clustering algorithm for diagnosing diseases. The work incorporates the classes of heart disease utilizing K-Nearest Neighbor and Naive Bayes. In this work, we have analyzed the use of K-Nearest Neighbor and Naive Bayes with different normalization techniques. The dataset utilized is the Cleveland heart disease from UCI machine learning repository. Our proposed works analyze the performance of K-Nearest Neighbor and Naive Bayes classification. The results prove that Naive Bayes classification gives better accuracy for diagnosing heart disease.
A.Amutha, M.Priya
In clustering technique, the hard clustering membership values and overlapping concept could not be identified along with non-convex problem. The proposed algorithm uses soft clustering to combine both Laplacians and multiple kernels for clustering analysis. The algorithm is formulated on a Rayleigh quotient objective function. The bi-level optimization is an alternating minimization procedure; it is used to convert the hard clustering to soft clustering. The kernels and Laplacians co-efficient can be optimized automatically by using the methods semi-infinite programming and quadratic constraint quadratic programming .The kernel Laplacians algorithm uses to control the overlapping.
S.Kokila, G.Pramela
In a computer network the transmission of data is based on the routing protocol which selects the best routes between any two nodes. Different types of routing protocols are applied to specific network environment. Three typical types of routing protocol are chosen as the simulation samples: RIP, OSPF and EIGRP. RIP (Routing Information Protocol) is one of the oldest routing protocols still in service. Hop count is the metric that RIP uses and the hop limit limits the network size that RIP can support. OSPF is the most widely used IGP large enterprise networks. OSPF is based on the Shortest Path First algorithm which is used to calculate the shortest path to each node. EIGRP (Enhanced Interior Gateway Routing Protocol) is Cisco's proprietary routing protocol based on Diffusing Update Algorithm. EIGRP has the fastest router convergence among the three protocols are testing. The main aim to analyze the performance of the three protocols such as their router convergence, duration and end-to-end delay. In this concept we are going to use OPNET to simulate RIP, OSPF and EIGRP in order to compare their attributes and performance. According to the convergence we want to find out which protocols are suitable for different sizes and types of network.
S.Jeevitha, M.Bhuvaneswari
Multimedia data mining is the process of identifying interesting patterns from multimedia data like audio, video, image and text which are not accessible by basic queries and associated results. It mines valuable knowledge and high level multimedia information from large database system and it includes pattern discovery, rule extraction and acquiring knowledge from large multimedia database. Multimedia data mining techniques are used to extract knowledge from multimedia database. In this paper, we focuses on various MDM techniques which perform clustering, classification, sequence pattern mining, association rule mining and result visualization. It is a survey paper on the problems and solution of multimedia data mining that involves the basic concepts, various application areas, techniques, and approaches of MDM. By analyzing this large amount of multimedia data to extract useful knowledge is one of the challenges that has opened the opportunity for research in MDM.
V.Meena Gomathy, M.Lalithambigai
The medical query related websites are developing in recent years and a large number of patients and doctors are involved. The valuable information from these medical query websites can benefit patients, doctors and the society. It has been a difficult process that to extract medical knowledge from the noisy question-answer pairs and filter out unrelated or even incorrect information. Facing the problem of getting information generated on the medical query websites every day, it is unrealistic to fulfill this task via supervised method due to expensive annotation cost. In this paper, it is to be surveyed a Medical Knowledge Extraction (MKE) System that automatically provides high quality knowledge extracted from the noisy question-answer pairs and also estimate doctor’s expertise who gives answers on these query websites. The MKE system is a truth discovery framework to estimate trustworthiness of answers and doctor expertise from the data. This further handle three unique challenges in medical knowledge extraction tasks as: representation of noisy input, multiple linked truths and the long-tail phenomenon in the data. The MKE system is applied on real-world datasets crawled from “icliniq.com”, one of the most popular medical query related websites. Both quantitative evaluation and case studies demonstrate that the proposed MKE system can successfully provide useful medical knowledge and accurate doctor expertise. We further demonstrate a real-world application: “Care for You “, which can automatically give patients suggestions to their questions.
L.Sankara Maheswari, V.Malathi
Propelled innovation in remote correspondence cleared route to the advancement of ease, low control and multifunctional sensor nodes in remote sensor systems. The plan of Wireless sensor system is impacted by components like versatility, energy utilization, and environment and so on. The vast majority of the energy is spent on correspondence purposes. Energy protection is in this way a predominant figure WSNs. Directing system determination is critical for appropriate conveyance of packets. Continuous research points in broadening system lifetime by outlining conventions that requires less energy amid correspondence. An energy gathering WSN is an answer against the waste of energy in battery fueled systems since reestablishment of vitality is excessively costly. Energy gathering make utilization of hubs that can collect energy from nature. The energy of every hub has its breaking point and can't be supplanted or energized. All parts of WSNs must be an energy productive segment, equipment segment as well as programming segment. Energy proficient steering convention can draw out the systems lifetime. Responsive WSNs is tended to in this work. A convention utilizing static bunching method with group head choice in light of most extreme remaining energy is proposed. Reenactment is performed to show the execution of the proposed convention. It is demonstrated that the proposed convention can drag out the system lifetime superior to that of the customary conventions.
K.Srishanmathi, D.Umamaheswari
Digital Image Processing deals with developing a digital system that performs operations on a digital image. Image processing involved various processes as segmentation, classification, recognition and restoration. Segmentation plays a vital role in image processing. Segmentation is a technique that extracts the interested region from the original image. Several algorithms and techniques have been buildup for image segmentation. Among the techniques of segmentation, Edge detection is used as the base of another segmentation technique. Edge detection techniques alter the image into edge image using the modification of gray tones in the images. Among the different edge detection techniques, canny edge detection is the best one to detect the wide range of edges in images. In image segmentation, K-means algorithm classifies the image into group of classes (cluster) based on their distance from one to another. In this paper, the techniques of image segmentation; canny edge detector and K-means algorithm are discussed.
Dr V. Krishnapriya, C.N. Indhu Mathi
By the virtue of enhancing technologies, all the devices in the home can be connected. Owing to the ubiquitous availability of WiFi, all the appliances in the home-based environment can be connected through a common gateway.WiFi is choosen as the mode of communication.This effectively improves the comfort, indoor security, and cost saving at the home. Remotely monitoring and controlling the home appliances in a case of power consumption and network bandwidth which becomes a major role and concern. we need a low power device that transmits messages through a less verbose protocol. This paper presents a overview of Message Queuing Telemetry Transport (MQTT) Protocol which is used for home based technologies.
R. Subha Sree, V. Malathi
We present a novel image stitching approach, which can produce visually plausible panoramic images with the input taken from different viewpoints. Unlike previous methods, our approach allows wide baselines between images and non-planar scene structures. Instead of 3D reconstruction, we design a mesh- based framework to optimize alignment and regularity in 2D. By solving a global objective function consisting of alignment and a set of prior constraints, we construct panoramic images, which are locally as perspective as possible and yet nearly orthogonal in the global view. We improve composition and achieve good performance on misaligned area. Experimental results on challenging data demonstrate the effectiveness of the proposed method. In proposed method that combined Harris with SIFT, using the Harris algorithm with adaptive threshold to extract the corners and the SIFT descriptor to make the registration, which generated a 360-degree panoramic image quickly. And there are many panoramic image mosaic software, such as Photo Stitch, Panorama Maker, PixMaker and so on. The key of constructing a measurable aerial panorama is obtain the position and poseur of the aerial panoramic image. Although these software are easy for us to generate panoramic image, the geographical location information of the image can't be obtained, and the problem leads the aerial panoramic image can't be measured and located. The algorithm of panoramic images and multi-view oblique images stitching is needed so that we can get the relation between them, which helps us locate the position of panoramic images.
S.Karthigai, Dr.K.Meenakshi Sundaram
Both men and women Cancer is the most important cause of death. The early detection of cancer can be helpful in curing the disease completely. So the requirement of techniques to detect the occurrence of cancer nodule in early stage is increasing. A disease that is commonly misdiagnosed is lung cancer. Earlier diagnosis of Lung Cancer saves enormous lives, failing which may lead to other severe problems causing sudden fatal end. Its cure rate and prediction depends mainly on the early detection and diagnosis of the disease. One of the most common forms to analyze the cancer in early stage is the Knowledge discovery in data mining have applications in ADALINE is Adaptive Linear Neuron Network domain. The ability of a net to learn a new pattern equally well at any state of learning is called plasticity. Discovery of hidden patterns and relationships often goes unexploited. Using generic lung cancer symptoms such as age, sex, Wheezing, Shortness of breath, Pain in shoulder, chest, arm, it can predict the likelihood of patients getting a lung cancer disease. Aim of the paper is to propose a model for early detection and correct diagnosis of the disease which will help the doctor in saving the life of the patient.
M.Kavitha, Jasmin Thomas
Rather being limited to pure academic research, Machine Learning Techniques and their practical applications is increasing their penetration into the main stream business processes. In business, correct decisions at the right time are quite important and have significant cost and risk implications. Decision Support Systems were developed for this where they assist and support key people to make business or organizational decisions. In this paper, we are discussing on the practical application of Machine Learning Techniques to predict the Employee Retention and optimize the hiring plan of an organization serving it as a decision support function for the Human Resource Management Systems. We have chosen Decision Forest and Artificial Neural Network as the algorithms and aim to discover the employee retention pattern in an organization.
N.Senthil Kumar, R.Umagandhi
Wireless Sensor Networks (WSNs) is a collection of sensor nodes with capability of sensing various types of environmental and physical conditions. The network is composed with an individual number of nodes placed in a wide area and communicates through radio interface. The main object of the WSN is to collect the data from the environment So, different types of routing protocols have been designed to manage various types of routing, power management, data distribution and QoS for WSNs. Routing protocols in WSNs is responsible for maintaining the routes in the network and have to ensure the reliable single-hop or multi-hop communication under the conditions. In this paper, we focused Flat Based network structure routing protocols for WSN and compare their strengths and limitations.
R. Mallika, V. Ramakrishnan
Data mining is a logical process that is used to search through large amount of data in order to find useful data. The objective of this system is to find patterns that were previously unknown. Once these patterns are found they can further be used to make certain decisions for development of their businesses. Various techniques like Classification, Clustering, Regression etc., are used for knowledge discovery from databases. Clustering analysis is one of the main analytical techniques in data mining. Clustering is the grouping together of similar data items into clusters. This paper discusses the various types of clustering algorithms and also analyse advantages and disadvantages of clustering algorithm based on various environments.
S.Renuga
Diseases in plants cause major production and economic losses as well as reduction in both quality and quantity of farming products. Disease management is a challenging task. Mostly diseases are seen on the leaves or stems of the plant. Precise quantification of these visually observed diseases, pests, traits has not studied yet because of the complication of image patterns. Hence there has been increasing demand for more specific and sophisticated image pattern understanding. This work presents a method for identifying plant leaf disease based on color. Agrarians are suffering from the issue rising from different types of plant leaf diseases. Sometimes biologists are also unable to identify the disease that leads to need of identification of right type of disease. First the input image is pre-processed. Then input image of leaves is converted as Red Green Blue (RGB) to Hue Intensity Saturation (HIS) or Lab color space. Then leaf disease segmentation is done using Hierarchical clustering. After segmentation the mostly green color pixels are covered based on specific threshold values. The Support Vector Machine (SVM) and Neural Network (NN) is trained for classification. The goal of this study is to provide, different identification techniques for plant leaf. Some automatic technique is beneficial as it reduces a large work of monitoring in big farms of crops, and at very early stage itself it detects the symptoms of diseases i.e. when they appear on plant leaves.
S.Saranya, P.Banumathi
Mining infrequent Itemset is fundamental method for mining association rules as well as for many other frequent Itemset mining tasks. In existing methods for Mining frequent and infrequent has been implemented using FP growth or Apriori algorithm. Also numerous experimental results have demonstrated that these techniques are scalable to the mining process. In this paper, we present a novel Transaction Mapping Technique for filtering the association rules for infrequent Itemset in the transaction dataset. Proposed technique produces improved performance for sparse data items. Furthermore, we present a open and closed Itemset extraction rules using optimization techniques. The experimental results prove that proposed technique highly scalable and consume less memory compared to the state of art techniques.
M.Kavitha, R.Baby
Clustering is a relevant task in mining data streams, which merge similar items in a cluster. Many clustering approaches are introduced in recent years for data streams that are based on density which provides protection against anomalies. Micro-clustering is a technique in stream clustering that stores the compact information of the data objects in data streams. Micro-cluster is a temporal extension of the cluster feature, which compresses the data effectively. The proposed system works a without limitation based algorithm with the purpose of automatically adapts to the speed of the information stream. It makes greatest use of the time available under the present constraints to provide a clustering of the objects seen awake to that point. The proposed approach incorporates the age of the objects to reveal the greater importance of more modern data. In efficient and effective handling, here we introduce the Decision Tree based DBSTREAM micro cluster (DTDBS), a compact as well as self-adaptive index formation for maintaining stream summaries. We present solutions to handle very fast streams at some stage in aggregation mechanisms and propose novel descent strategies so as to improve the clustering effect on slower streams as long as time permits. Our experiments show that our approach is capable of conduct a multitude of different stream characteristics designed for accurate and scalable anytime stream clustering.
S.C Divya
Skin dermis is mostly found in humans, animals and plants. A skin defects is a particular kind of infection caused by bacteria or virus. These diseases like alopecia, ringworm, yeast infection, brown spot, allergies, eczema etc. have various perilous effects on the skin and keep on spreading over time. It becomes imperative to diagnose these defects at their pioneer stage to control it from spreading. These diseases are sanctioned by using many technologies such as image processing, data mining, artificial neural network (ANN) etc. Recently, image processing has played a primary role in this area of research and has generally used for the detection of skin diseases. In this paper we investigate two methods for describing the contents of images. The first one characterizes images by colour spacing, while the second is based on histogram and histogram equalization approach and feature extraction etc. are part of image processing and are used to describe the part affected by disease, the form of affected area, its afflicted area colour etc. An exhaustive learning of skin disease diagnosis systems are done in this paper, with different methods and their performances. These techniques are carried out with an experiment by using MATLAB software. It is found that further modifications are needed to produce better performance in searching images.
Dr. R. Hemalatha, A.Babyshalini
Chronic Obstructive Pulmonary Disease (COPD) is the important cause of death in the world. Some most common chronic diseases are COPD, diabetes, cardiovascular disease and chronic respiratory disease. Early detection can save the life and survivability of the patients. In this paper propose model give a solution to predict chronic diseases. In this paper proposes a novel approach of applying the Ant Colony Optimization technique (ACO) for extracting the Association Rules (AR) from the database to detect COPD. This algorithm is broadly divided into three parts, in the first part, accept the data set of chronic symptoms which is a generalized way for creating the patterns for Chronic diseases Framework, and in the second part, find the relevant data from the patterns. It can choose the frequent symptoms only by using the support count value. The pheromone value which the support of the pattern of COPD symptoms. Subsequently by outcome analysis, prove the effectiveness of this algorithm. The focus of this paper is to provide specific information about chronic diseases for public. Data mining is the process of discovering interesting patterns and knowledge from large amounts of data. Preventive Health care knowledge is essential for clinical and administrative decision making.
K.Devipriya, L.Subathra Devi
Early detection of skin cancer has the potential to reduce mortality and morbidity. In this world near about 1/7th of total world population suffer from some sort of skin disorder. This paper presents two hybrid techniques for the classification of the skin images to predict it if exists. The proposed hybrid techniques consist of three stages, namely, feature extraction, dimensionality reduction, and classification. In the first stage, we have obtained the features related with images using discrete wavelet transformation. In the second stage, the features of skin images have been reduced using principle component analysis to the more essential features. In the classification stage, two classifiers based on supervised machine learning have been developed. There are six different categories of skin diseases which shares somewhat same features. So for the classification of these diseases Bayes net a Bayesian technique along with feature selection has been used in this study. The first classifier based on feed forward back-propagation artificial neural network and the second classifier based on k-nearest neighbor. The classifiers have been used to classify subjects as normal or abnormal skin cancer images. A classification with a success of 95%and 97.5% has been obtained by the two proposed classifiers and respectively. This result shows that the proposed hybrid techniques are robust and effective. Using Expert System it exhibits the diagnosis of skin disease identification accuracy of 85% for Eczema, 95% for Impetigo and 85% for Melanoma.
Dr.R.Hemalatha, G.Kalaivani
The present Wi-Fi usage is frequently statically jumped to the position of mobile strategy. While this “on-the-spot” Wi-Fi off-load is tranquil effectual, current study propose that one can further expand the benefit of Wi-Fi way in if we let delay tolerant between the network connections. Current versatile applications (e.g., online video players, podcast applications, and reinforcement applications) are progressively more data transfer capacity hungry. Sadly, because of constrained radio assets and foundation limit, it is vague whether cell ISPs can take care of quickly expanding movement demand. Be that as it may, Wi-Fi can give higher data transfer capacity at a lower cost than cell organizes. D2TP is a vehicle layer convention for versatile applications, giving TCP-like, solid information move in stationary situations. It conceals arrange disturbances and permits postpones when a cell phone is progressing. The key empowering influence for D2TP disturbance resistance is in the partition of an association from its system connection. 3G Cellular systems are directly confronting extreme movement over-burdening issue caused because of inordinate movement requests from versatile clients. The current systems must be giving irregular network to the clients. Non-unimportant deferral may come about by using them for movement offloading. This postponement won't make versatile clients fulfill. Offloading part through Delay Tolerant Networks and Wi-Fi hotspots is a most appropriate arrangement. Keeping versatile clients fulfillment in see, there is have to give a motivating force system to use postpone resistance for cell movement offloading. Proposed Research has distinguished that by limiting the motivator cost, clients with high postpone resilience also, extensive offloading potential ought to be given higher need. In this research, we examined a motivating which works on turn around sell off in which client proactively express their postponement resistance through offers accommodation. We additionally examined how both DTN and WiFi hotspots can predicts the offloading capability of the clients.
M.Kavitha, K.Hemapriya
Wireless sensor networks (WSNs) are increasingly being disposed in security-critical applications. Because of their inherent resource-constrained characteristics, they are prone to various security attacks, and a black hole attack is a type of attack that seriously affects data collection. To conquer that challenge in the existing system, an active detection-based security and trust routing scheme named ActiveTrust is introduced for WSNs. ActiveTrust can extremely increase the data route success probability and across against black hole attacks and can optimize network lifetime. However in the existing system, the computation overhead is high which finding the trustable detection routes in case of presence of higher nodal density where the request needs to be traversed through all nodes, The existing work only concentrates on detection of black hole attacks which doesn’t consider the grey hole attacks present in the network. This is resolved in the proposed system by concentrating secured and trustable routing which can improve the overall performance of the wireless sensor network with improved packet delivery ratio. This is done by filtering out the packets with less trust value to avoid the unwanted transmission of control packets. This is done by introducing the new method namely Trust based Packet Filtering (TPF) where the packets with less trust values would not be considered for the further packet transmission. In addition to that Trust and Density aware Clustering (TDC) method is introduced to avoid the more computation overhead. And also in this research work, security is enhanced by finding the grey hole attacks along with black hole attacks. Here adversary nodes are identified based on the information gathered from the neighbor nodes by using which more optimal cluster head is selected. Thus the optimal and secured routing can be ensured in the wireless sensor network which would increase the WSN contribution.
D.Kaviya vikashini, P.Sangeethaa, K.Umaiyal
The emergence of new types of crime as well as the commission of traditional crimes by means of new technologies is called as “Cyber-crime”. Computer crime is defined as criminal activity involving an information technology infrastructures including illegal access like unauthorized access, illegal interception is done by technical means of computer data from or within a computer system, data interference like illegal damaging, deletion, deterioration, modification or suppression of computer data, systems interference like interfering with the functioning of a computer system by inputting, transmitting, damaging, misuse of devices, forgery is also called as ID theft . Cybercrime is a relatively new phenomenon. Services such as telecommunications, banking and finance, transportation, electrical energy, water supply, emergency services, and government operations rely completely on computers for control, management, and interaction among themselves. Cybercrime would be impossible without the Internet. . Other than computer viruses, specific crimes dealing with computers and networks (such as hacking) and the facilitation of traditional crime through the use of computers (child pornography, hate crimes, telemarketing/Internet fraud).
Dr.C.Kalaiselvi, G.Arunasenbagam
Key management in the unprepared network is an exigent problem concerning the security of the group communication. There are three categories for classifying Group key management protocols; centralized, decentralized with distributed. in establishing key management protocol, suitable solution can be provide to services like verification, data reliability and data confidentiality. This paper deals by an approach for designing and analyzing the region-based Group key management protocols in support of scalable and reconfigurable group key management in Mobile unprepared Networks (MANETs). The main problem of centralized key management protocols is a propos data security planned group communication. To rise above this problem a narrative approach for key management into Region based MANET is proposed. The Group key organization comprises creating and distributing a common covert for all the group members. However, key management for a large and dynamic group be a difficult problem because of scalability and security. adjustment of membership requirements the group key to be refreshed to make sure backward and forward secrecy. In this paper, a Robust and Efficient Group Key(REGK) management scheme be proposed for Region based MANETs. The proposed method is as well effective in defending against many complicated attacks such as rejection of service (DoS) attack. In order toward preserve the security, the region-based group key management protocols deal with stranger attacks in MANETs. The experimental results compares the calculation cost moreover time for the existing and proposed approach and the results illustrate so as to the proposed approach outperforms the existing method through lesser calculation cost and time.
Dr.S.Radhimeenakshi, K.Latha
Stock market analysis have major impact on economy condition of any country as well as on the global economy. Stock activities estimating is a high demand for stock customers. This stock estimating is a challenging issue. Hence, we must a need to develop application that is capable to exactly predict directions of stock price movement. Our paper suggest a data mining technique to model relationship among company stock with other companies stocks. It is conventional that selected rules can be of a help to guess future stock market prices movements with significant level of accuracy.
Dr.R.Hemalatha, R.Indhumathi
In the fast development world, the energy efficiency of wireless networking protocols becomes a concern for many stakeholders. The wireless networking protocol can also be used for gathering the underwater data and routing in the data using the source and destination node. The routing plays a main role for collecting data in networks, in such a way it is required to play the networking in underwater, and a way to localize the device nodes makes an incredibly in cost and error prone situation. The packets are routed from one to another node, where hop to hop network is measured. The Underwater Wireless Sensor Network (UWSN) is a model to discover underwater environment. The uniqueness of Mobile Underwater sensor networks has a problem in low communication bandwidth, large transmission delay and mobility of nodes. In this proposed research, a new efficient routing protocol is to be introduced, such protocol gives multipath utilizing method from source to destination, the series with sub paths are already taken into account by its introduced protocol and also helps to calculate the neighboring node from source to destination nodes. In such a way, relay nodes can be calculated from nodes using transmission delays. Through this work, it send the packets very secure by multipath communication in underwater sensor networks, and secure communication helps to send or receive the packets safely.
V.Shanu, S.Vydehi
Medical data mining is an active research area in the present scenario. Medical data analysis and disease diagnosis have great impact on several medical systems. This includes heart disease prediction, diabetes detection, cancer and other type of health disorders. Data mining is the optimal choice to accomplish those processes in medical dataset. Optimal clustering in health care dataset is an important task due to its huge dimensionality. Medical data clustering emerges with numerous research challenges like clustering accuracy, delay, and minimizing intra cluster distance. In this paper, we propose a novel technique to perform optimal clustering on two different medical datasets heart disease and liver disease. To improve the cluster performance and accuracy, an optimization algorithm is used. The proposed system increases the cluster quality by deploying a hybrid technique which combines weighted fitness firefly (WFF) and Modified BAT (MBAT) Optimization Techniques. The modified BAT (MBAT) technique reduces the time utilization and WFF finds the optimal feature for cluster. Instead of random move of firefly, the optimal movements are identified and performed first. The MBAT is mainly used to reduce the multimodal optimization problems by applying the hybridization techniques. The results and experiments generated. And the proposed system shows the improvement on accuracy, specificity, and consistency etc.
Dr.K.Rajeswari, I.Meena
Antivirus software is tools purpose for decision and stopping cruel and unnecessary files. However, the long term effect of traditional host based antivirus is questionable. Antivirus software fails to detect many modern threats and its increasing complexity has resulted in vulnerabilities that are being exploited by malware. The promote a new model for malware detection on end hosts based on providing antivirus as an in-cloud network service. This Antivirus enables identification of malicious and unwanted software by multiple detection engines Respectively, Our Enhanced Identity Based Signature with Malware detection approach of moving the detection of malicious users into the cloud is aligned with a strong trend anent moving services from end host and monolithic servers into the cloud. This approach provides several important benefits including better detection of malicious software, enhanced forensics capabilities and improved deploy ability. Enhanced Identity Based Signature with Malware detection in cloud computing includes a lightweight, cross-Storage host agent and a network service. Here a static signature verification system based on the concept of local stability. Stable regions are detected in the signatures, during the enrolling phase, and are considered to the those regions affected by low variants of features among the training set. The stability evaluation is based on the Hamming distance. Stable regions are successively used for verification in the running phase. A region oriented verification strategy is considered, based on a well-defined similarity measure which takes into account the variability in signing of the writer. The results, carried out on signatures from the GPDS database, expose the viability of proposed approach.
P.Dhivyabharathi, V.Priya
People may share a transmitting and storing millions of images each and every second. For transmission and storage of large dataset image compression technique is an essential approach. Although, data compression is typically done to avoid the more memory space and to enhance capacity of the storage devices. Consequently, to order of perfect, image compression algorithm is very high which can be used to make something less in size the resources usage. This technique is presented the study of various lossless compression and lossy compression techniques.
B. Nandhini, M. Praveena
The service oriented ad hoc networks are consolidated with network service providers and service requestors. This kind of network doesn’t like to have malicious nodes which may collude to maximize their own gain and even monopoly service. Trust calculation for finding malicious service requestor as well as service providers are much complicated. The trust management suffers from several attacks and issues such as bad mouthing attack, self promotion, ballot stuffing and opportunistic service attacks etc., This paper provides the survey of various techniques and methods involved with the trust calculation and multi objective optimization in service oriented ad hoc networks.
M.Shanthakumar, S.Janarthanam, S.Sukumaran
The development of detectable points is important in image processing the effective fuzzy oriented saliency detection make advances in image contents. Usually the take the role normal are identified by bilateral filters and retaining the local features. Fuzzy oriented method with patterns between the center pixel and its surrounding neighbours in two dimensional local region proposed robust fuzzy salient region extraction algorithm (FSE) encodes the spatial relation between any pair of neighbours in a local region along the directions for the center pixel in an image. The experiments carried out for proving the worth of proposed method on two different types of benchmark databases and the new metrics Directional based Gaussian weight (DGW) and Relative Normal Distance (RND) has been proposed in this paper. The comparative studies based on absolute data from two publicly accessible databases show that the proposed method usually outperforms both qualitatively and quantitatively for saliency estimation analysis and understanding. It can be further extended with implementation of new retrieval techniques with different parameters which improves more retrieval efficiency and performance integrity.
G.Divya, R.Nithyaananthi
In this paper we propose a pattern mining technique to study event detection representation from difficult multivariate temporal data, such as electronic health reports. Pattern recognition is seen as a main challenge within the field of data mining and knowledge discovery. In this paper, we propose a comprehensive data mining framework for event detection DP miner, which functions in a distributed and parallel manner (data in a partitioned database processed by one or more sensor processors) and is able to extract a pattern of sensors that may have event information with a low communication cost. To achieve this, we introduce a new sensor behavioral pattern mining technique called sequential data mining. The task of sequential pattern mining is a data mining task specialized for analyzing sequential data, to discover sequential patterns. More specifically, it consists of discovering interesting subsequences in a set of series, where the interestingness of a subsequence can be calculated in terms of a range of criteria such as its occurrence frequency, length, and profit. In order pattern mining has several real-life applications due to the actuality that data is naturally encoded as series of symbols in many fields such as bioinformatics, e-learning, market basket analysis, texts, and web page click-stream analysis. An Apriori algorithm has been proposed for data preparation, to generate sequential sensor patterns. Evaluation results show better trade-off between Sequential data mining and Differential data mining. An analysis for communication cost is also evaluated here.
Dr.K.Rajeswari, Meenaatchi.S.M
Preventing the human being from Diabetes is a very challenging role towards the routine life for everybody in the World. Patients affected with diabetic with the high level of blood sugar have to be monitored in regular way. Consequently, it will also affect the organ of the human body as much as faster, when the range gets exists. At any stage of diabetes value penetrating, Some Symbolic Identification can support as to prevent us to warn the stages to safe-guard. Early stage Symptoms Identification supports us to protect the Particular organ in many ways of Medical Approaches. Diabetic Neuropathies will affect the organs like feet, toes, heart, arms, legs, urine tract and soon. A Quick Decision Repository is applied to fine tune the organ in early stage.
C.Mohanapriya, C.Hemapriya
The journal investigates the following fundamental question how fast can information be collected from a wireless sensor network organized as tree? To address this, a number of different techniques using realistic simulation models under the many to one communication paradigm known as converge-cast are evaluated. Time scheduling on a single frequency channel with the plan of minimizing the number of time slots needed (schedule length) to whole a converge-cast is considered. In present system they have present Kautz graph based method, in our proposed system we use Tree based wireless sensor network system by, scheduling with transmission power control is combined to mitigate the effects of interference, and show that while power control helps in minimizing the schedule length under a single frequency, scheduling transmissions using multiple frequencies is more efficient. Converge-cast, namely the collection of data from a set of sensors toward a common sink over a tree based routing topology, is a fundamental operation in wireless sensor networks (WSN). In this journal, consider a TDMA framework and design polynomial-time heuristics to reduce the schedule length for both types of converge-cast. It also locates lower bounds on the possible schedule lengths and compares the performance of our heuristics with these bounds. Lower bounds on the schedule length are given when interference is entirely eliminated, and propose algorithms that achieve these bounds. Then, the information collection rate no longer remains restricted by interference but by the topology of the routing tree. Fast data collection with the aim to reduce the schedule length for aggregated converge-cast. It examines the impact of transmission power control and multiple frequency channels on the schedule length, where the proposed constant factor and logarithmic approximation algorithms on geometric networks (disk graphs).
D.Hemavathi
Document clustering intends to automatically group associated documents into clusters. This proposed work presents a new spectral clustering technique called Correlation Preserving Indexing (CPI), which is performed in the correlation similarity measure space. In this framework, the documents are projected into a low dimensional semantic space in which the correlations between the documents in the local patches are maximized while the correlations between the documents outside these patches are minimized simultaneously. Since the intrinsic geometrical structure of the document space is often embedded in the likeness among the documents, correlation as a similarity determine is more appropriate for detecting the intrinsic geometrical structure of the document space than Euclidean distance. Accordingly, the proposed CPI technique can effectively find out the intrinsic structures embedded in high-dimensional document space. In addition, the proposed work considers the major variation among a traditional dissimilarity/similarity measures and is that the former uses only a single viewpoint, which is the origin, while the latter make use of many different viewpoints, which are objects assumed to not be in the same cluster with the two objects being measured. More informative assessment of similarity could be achieved by using multiple viewpoints.
M. Sathyapriya, K.A. Poornima
In today's world, the most accepted payment mode is credit card for online transactions which provides cashless shopping at every corner across the world. It is the most suitable way to do online shopping, paying bills, and performing other related tasks. Hence risk of fraud transactions using credit card has also been increasing. In the prevailing credit card fraud detection processing system, fraudulent transaction will be detected after transaction is done. Hidden Markov Model is one of the statistical tools for engineers and scientists to solve various problems. Credit card frauds can be detected using hidden markov model during online transactions. Hidden markov model aids to obtain a high fraud transaction coverage combined with low false alarm rate, thus providing a better and convenient way to detect frauds. Using hidden markov model, customer’s pattern is analysed and any deviation from the regular pattern is considered to be a fraudulent transaction. So, hidden markov model is initially trained with the normal behaviour of a cardholder. If an incoming online card transaction is not accepted by the trained HMM with high probability, it is considered to be fraudulent. At the same time, the algorithm tries to ensure that genuine transactions are not rejected. In this paper, the sequence of operations in online card transaction processing is modelled using Hidden Markov Model (HMM) to detect fraudulent transactions.
Dr. N. Sasirekha, A.Shanthi Sona
Internet has changed and improved the way of working in organizations and businesses, at the same time this large network also opened doors for attackers as new attacks are emerging day by day. To protect the organizations and systems from these attacks, network security comes into action. The need for computer intrusion forensics arises from the alarming increase in the number of computer crimes that are committed annually. After a computer system has been breached and an intrusion has been detected, there is a need for a computer forensics investigation to follow. The goal of this paper is to explain the advantages and disadvantages of computer intrusion forensics. The paper will look at how intrusion detection systems can be used as a starting point to a computer forensics investigation. Also, the ways to preserve and recover data during a computer forensics investigation will be explored. A discussion of how some of various software tools that are used in a computer forensics investigation will be included. Last, the paper will explore ways that an intrusion detection system can be used in correspondence with computer forensics.
P. Banumathi, I. Buvana
As the Internet takes more and more central role in our communications infrastructure, the slow convergence of routing protocols when a network failure becomes a growing drawback. To assure quick recovery from link and node failures in. To assure recovery of packet from the node failures, the networks have the present scheme of recover called multiple routing schemes, in the present research they use Multi slot max clique method to handle number of packets losses in the network, in present they handle unique coverage problem for communication and they handle with optimal solution strategy technique. Our proposed scheme is based on multiple routing from the source to destination but they have assume that the hop-by-hop forwarding is better than present system, because the routing information is stored in routers and it allows packets to forward from one to another hop, thus the packet forwarding link is immediately detection of failure using our approach ACK packets, and the algorithm here followed in Optimal Packet Forward (OPF) as, it has many security and inbuilt schemes to forward the packets.
V.Ramya, S.Thavamani
Outsourced Electronic Health Record to the cloud for the high quality retrieval and storage service has experienced large security violation. However it may lead to leakage of sensitive information of the patient. In order to protect the data leakage, efficient secure data sharing models has been proposed in the literature. In this paper, we analyze the several secure data sharing gateways in the cloud utilizing searchable encryption and proxy re encryption. The major primitive is public key encryption scheme with keyword search which enables the data users to search on the encrypted information without decrypting it and proxy re- encryption can be introduced to conjunctive keyword search in terms of the time enabled proxy Re-encryption model. It enables the data owner to delegate the access rights to data user to operate several search keywords which is considered as conjunctive keywords on their records within the specified time and providing resistance against guessing attacks. Time based delegation can also leads to security violation issue. To handle the implication on this study, we propose a novel hybrid constraint based re encryption (HCRE), multiple constraint such as user category, time and Specific intension of the user are taken as a key to re encrypt the data from security violation. The HCRE can improve the performance in terms of security and computation cost.
N.Revathi, M.Madhangiri
It is a very difficult for a passenger to find out an optimal route to travel in a metropolis. A novel approach of search the optimal routes based on date mining technology is presented in this paper. In the approach, we first use the Ant Colony-based Data Miner algorithm to search and to get the candidate set of the optimal bus-routes, and then use data mining method to mine the optimal lines which is hidden in the candidate set. The goal of Ant-Miner is to extract classification rules from data. Ants often locate the shortest path among a food basis and the nest of the colony without utilizing visual information. In order to replace information about which path should be go after, ants communicate with each other by means of a chemical substance called pheromone. The algorithm is inspired by both research on the performance of real ant colonies and a few data mining concepts and principles. The mutual use of ACO and DM (the utilization of ACO algorithms for DM tasks) is a very capable direction. In this chapter, we evaluate ACO, DM, categorization and Clustering (two of the most accepted DM tasks) and focus on the use of ACO for Classification and Clustering. Furthermore, we briefly present associated applications and examples and outline possible future trends of this promising collaborative use of methods.
D.Yamuna, Dr.N.Sasirekha
Software testing is some action meant by evaluating an attribute and determining that program or system meets its required results. This is significant accomplishment in software development. Test case selection is a critical action in testing because the number of automatically generated test cases is regularly huge and probably unfeasible. As well, a large number of test cases are unnecessary. It trains similar features of the application and they are capable of uncovering a similar set of faults. Data mining finds similar patterns in test cases which helped us in finding out redundancy incorporated by automatic generated test cases. We proposed a methodology based on data mining by which we can significantly reduce the test suite. The paper aimed to selecting the fewer related test cases at the same time as providing the best possible model from which test cases are generated by using data mining techniques.
R.Narmatha, D.Rajalakshmi
Big Data is a large-volume, complex, growing data sets with multiple or autonomous sources. This huge amount of data is generated by social media and networks, scientific instruments, mobile devices, sensor technology and networks. These data sets are able to manage, analyze, summarize, visualize and discover knowledge from the collected unstructured data in a timely and scalable manner is very complex task using usual data mining tools. Data Mining is an analytic process with great potential, designed to discover large amounts of data also known as “big data” and explore for consistent patterns and systematic links between variables and then to validate the findings by applying the detected patterns to form new subsets of data. This paper begins with a brief introduction to data mining, followed by the discussions of big data analytics and current status, Controversies and some challenges are also be presented.
Dr.C.Kalaiselvi, V.Mehalarajini
A major part of current research in data mining is the ground of medical diagnosis. In the present learning with the Breast cancer Wisconsin data sets, a feature selection algorithm Modified Support Vector Machine Feature Selection (MSVMFS) predicts together diagnosis and prognosis by comparing several data mining classification algorithms. In the proposed approach, in level one of feature selection, features are selected based on rough set among different starting values of feature reduction. In level two features are selected since the reduced set based on the Correlation Feature Selection (CFS). Experiments show the proposed method is valuable by comparing through others in terms of number of preferred features and classification performance.
L.Sudha
Routing algorithm is a major division in network layer for design the path to transmission the information. The network layer making route packets from source to destination for selecting a route through the network and in general more than one route is possible root. Based on performance criteria, route selection will be done and the finding the shortest route and select path that passes through the least number of nodes which results in the least number of hops per packet to perform this task to reach the goal. The Poisson distribution can be applied to systems with a large number of possible events, each of which is rare. How many such events will occur during a fixed time interval? Under the right circumstances, this is a random number with a Poisson distribution. The conventional definition of the Poisson distribution contains two terms that can easily overflow on computers: λk and k!. The fraction of λk to k! can also produce a rounding error which is very large compared to e−λ, and therefore give an erroneous result.
D.Kavya, V.Sathyavathy
This paper manages the outline of a bio-metric security framework based upon the unique mark and discourse detective technology. In the principal section there are the bio-metric security frameworks and an idea of a coordination of the both advances presented. At that point the unique mark innovation took after by the discourse sleuth technology is right away portrayed. There are examined some fundamental standards of each of the advances.
S.Divya, G.Maria Priscilla
Email Classification is one of the vital problems in the email management due to its impact on the usage. Despite of several applications like messengers such as watsup, kaizala and social media networks such as Facebook and twitter, importance of email was kept exploring . In order to increase the performance of management it has become mandatory to automate the classification of the email against relevant and irrelavent emails This paper investigates the classification algorithms used to classify the email in terms of feature selection methods like genetic algorithm, simulated annealing and principle component analysis. This study discussesabout the importance of the classification modelsagainst the accuracy and security measures. To handle the implication on this study, we propose a novel deep learning based classification algorithm named as EmailGrading. In this feature are extracted from the low level feature in order tomaintain the hierarchical representation. Also it disentangles the abstraction on the different layers in order to improve the performance in terms of labelling and accuracy.
V.Muthulakshmi, N.Sumathi
The large enormous data is outsourced to the cloud for scalable data storage. The outsourced data is encrypted due to privacy and confidentially concerns. However importance of the keyword search on the encrypted data motivates the use of the searchable encryption. In order to encourage the multiple keyword searches, many state of art of approaches in implemented. In this paper, we analyse the multi keyword search mechanisms on the encrypted data with inclusion of data updating. The searchable encryption on multiple keywords is employed using vector space model and Tf-idf concepts to generate the index and query for the multiple random keywords. The encrypted index structure is constructed for data files. During the search operation, the search mechanism integrates the trapdoor of the keywords with index information and returns the matched record to the user. In order to reduce the processing time to the large amount of the data request ranked keyword mechanism can be utilized. The state of approaches is only restricted to the conjunctive searches and it is supported only to exact matches of the keyword. To tackle this implication, we propose a fuzzy based multi keyword search based on LSH function using the hash function to generate index and query. The utilization of this model, misspelled keyword can also be hashed into the query vector. The Euclidean distance is used to capture the keyword similarity. The fuzzy based multi keyword search mechanism can improve both efficiency and accuracy.
S.Selvi, S.Nirmalajancy
Bit-Torrent is one of most famous peer to peer file distribution server, that distribute large files in peer to peer manner such as Books, High Definition videos, Movies, albums, software’s, Television serials, other than available in the Internet in the format of *.torrent files. It is also a Peer to Peer protocol, nowadays it is very popular software to receive the file from any server in the world. In Such case, it is world wide resource providing protocol through this peer to peer network, we can also download many number of useful files through internet. In all the networking services there have an intruder, in same way peer to peer have an high security threats nowadays. Such security threats are legally attacked, by many intruders through virus, worms, Trojans etc. Through this security threats many of them fear to use of Bit-Torrent protocol, through using of this software we defined the attack of threats to that protocol system users, so the user need to verify the downloaded content and discover if it is trusted or not. We need to stop the attackers who hamper the illegal file distribution in Bit-Torrent Protocol. In Existing research, they made Piece-Attack in Bit-Torrent protocol, in same way we need to prevent the attack and give the security of downloading to end users, the piece attack is an attack against the leeches in torrent networks, that is observed against attack has to be fully scaled. Not only the Piece attack, there having peer attack, fake-block attack, benevolent attack peers and also famous DDoS attack. Any attacks can to be prevent but in existing research they made only for piece-attack prevention. Proposed Research executes the prevention and secure based torrent files download from the peers, this is based on peer tier architecture. The architecture builds the peer in the manner when the receiver receives from the torrent files, the tier architecture is started from that leech peer and receiving peers are in the tier are made, from the tier the message is send to each and every peer to secure their content.c
M.Preethi, P.K.Mangaiyarkarasi
Knowledge mining means extracting the knowledge from the database. Software algorithms are implemented for extracting the knowledge from database. In data mining each algorithm has a different objective and to obtain meaningful and previously unknown patterns from large dataset is an emerging and challenging problem. The capacity of the hardware architecture is fixed. As number of candidate itemsets or the number of items in the database is larger than the hardware capacity. So That the items are loaded into the hardware separately, Due to this time complexity is more to load candidate itemsets or database items into the hardware is in proportion to the number of candidate itemsets multiplied by the number of items .in the database. Increase of candidate itemsets and a large database would create a performance blockage. We propose a Hash-based and Pipelined (HAPPI) architecture for hardware-enhanced association rule mining. In HAPPI architecture, we propose PHP algorithm.
R.Brindha, P.Anitha
Diabetes is a chronic disease that contributes to a significant portion of the healthcare expenditure for a nation as individuals with diabetes need continuous medical care. Currently in the healthcare industry different data mining techniques are used to mine the interesting pattern of disease using the statistical medical data with the help of different machine learning techniques. The proposed system assists doctor to predict disease correctly and the prediction makes patient and medical insurance are also get benefited. This research focuses on to diagnosis diabetes disease as it is a great threat to human life worldwide. The system uses the K-Nearest Neighbor (KNN) and ID3 Algorithms as supervised classification models. Finally, the proposed system calculates and compares the accuracy of ID3 and KNN and the experimental result demonstrates that the ID3 provides better accuracy for diagnosis diabetes. For the clinical database, the Pima Indians Dataset is used in this research.
Dr. R. Hemalatha, V.Pavithra
Underwater communication is a technique that is used to send and receive message under water, There have several ways to communicate, in underwater through hydrophones, communicators, but the hydrophones are only used for some distance to gather the messages from the underwater, in such a way the underwater communication has made some difficult propagation such as multi-path communication, it is varied from bandwidth allocation, signal interference, strong signal, long range communication etc. In Underwater communication there have only low data rate communication to the terrestrial communication, in such a way the electromagnetic waves plays a main role in under water communication through acoustic waves. Building of Low cost IoT based sensors are very easy to implement nowadays, in such a way Internet of Things technology plays a very important role to all devices related to computer, The process of underwater communication through one node to another node gives an more challenges such as high path loss, limited available bandwidth, limited battery capacity, high bit error rate etc. The problem given is that the underwater sensor need to balance the energy, limited battery capacity and high bit error rate to be reduce, by this BEAR (Balanced Energy Adaptive Routing) nodes to communicate directly, it increases number of packets dropped in nodes stay alive for the longer period for throughput, in such a way we need to less the rate of packets dropping and nodes to be stay alive for longer period. In our research, we have proposed Protocol named as UERP (Underwater Vector based Routing protocol, it plays as location based protocol, thus the protocol is designed to routing the packets in vector based movements, thus the here Packet Adaption algorithm plays a main role to forwarding policy method to save battery (energy) of devices.
Dr. R. Hemalatha, N.Ramya
Parkinson Disease (PD) occurs due to the loss of dopamine in the brains thalamic section that results in unconscious or oscillatory movement in the body. Normally Doctors diagnosis the PD disease clinically with their expertise and experience. But most of the time immoral diagnosis and treatment are reported. For this, patients need to take number of tests for diagnosis, but the time, these all tests still not sufficient to diagnosis Parkinson Disease effectively. This work emphasis on classify the severity of PD or idiopathic Parkinsonism. Firstly, this work is proposed to apply some data mining technique to select the best attributes (according to the posture datasets). In second step, selected attribute was trained using motor evaluation UPDRS data and the algorithm is applied for the best attribute. Unified Parkinson Disease Rating Scale (UPDRS), captures numerous aspects of PD that include Mentation, Behaviour and Mood, Activities of Daily Life (ADL), Motor Examination and Complications of Treatment. For discovering the best classifiers Support Vector Machine-Sequential Minimal Optimization (SVM-SMO) algorithm is used. The accuracy is improved using algorithm. Data mining models is to monitor the sequence of gait features, which may eventually lead to earlier diagnosis of rising neurological disease for healthcare community.
R.Soniya, Boopalan.S
Target version is declared for the android apps. At the point when keep running on gadgets with later Android forms, applications are executed in a similarity mode that endeavors to copy the conduct of the more established target rendition. This outline has genuine security results. Applications that goal out of date Android adjustments cripple basic security changes to the Android organize. We call the issue of uses concentrating on out of date Android frames the target irregularity issue. We separate a dataset of 1,232,696 free Android applications accumulated between May, 2012 and December, 2015 and exhibit that the target intermittence issue is an authentic stress over the entire application organic group and has not changed amazingly in a significant extended period of time. By and large, 93% of current applications center out of date arrange frames and have a mean oldness of 686 days; 79% of uses are starting at now obsolete on the day they are exchanged to the application store. Finally, we investigate seven security related changes to the Android arrange that are disabled in applications that goal out of date organize frames and exhibit that target brokenness hamstrings attempts to upgrade the security of Android applications.