A. Dharmarajan ¹, T. Velmurugan ²

Affiliations
1 Bharathiar University, Coimbatore - 641046, IN
2 Research Department of Computer Science, D. G. Vaishnav College, Chennai - 600106, Tamil Nadu, IN

Abstract

Objective: The objective of this research work is focused on the ethical cluster creation of lung cancer data and analyzed the performance of partition based algorithms. This research work would help the doctors to identify the stages of lung cancer and also enhances the medical care. This work is very convenient to avoid unnecessary biopsy. Methods: Lung Cancer is the form of cancer that has caused the most deaths in both men and women throughout the world. Most of the researchers analyzed the lung cancer dataset using algorithms to find the cluster among the small cell or non-small cell lung cancer in various stages. The very famous two partition based algorithms namely k-Means and FarthestFirst are implemented. A comparative analysis of clustering algorithms is also carried out using two different dataset. The performance of algorithms depends on the time taken to form the estimated clusters. Findings: The performance and cluster formation using the two various kinds of input dataset namely lc.arff, lc.csv are used. The output clusters depends upon the dataset type and algorithms related. The number of initial clusters is chosen by the user. The data points in each cluster are displayed by different colors. The computational complexity is calculated in milliseconds. The k-Means algorithm is efficient for clustering the lung cancer dataset with arff file format. The final outcome of this work is suitable to analyses the behavior of lung cancer in the department of oncology in cancer centers. Our findings are well fit for report preparation and treatment selection of the patients. Application: The created ethical cluster is used for support ingredient of the department of molecular oncology in cancer institution or centers. Ultimate goal of this research work is to find out which type of dataset and algorithm will be most suitable for analysis of lung cancer data.

Keywords

Cluster Analysis, Farthest First Algorithm, k-Means Algorithm, Performance Analysis

Full Text

   

C. Anuradha ¹, T. Velmurugan ²

Affiliations
1 Bharathiar University, Coimbatore, IN
2 Department of Computer Science, D.G.Vaishnav College, Chennai, IN

Abstract

Objectives: Data mining techniques are implemented in many organizations as a standard procedure for analyzing the large volume of available data, extracting useful information and knowledge to support the major decision-making processes. Data mining can be applied to wide variety of applications in the educational sector for the purpose of improving the performance of students as well as the status of the educational institutions. Educational data mining is rapidly developing as a key technique in the analysis of data generated in the educational domain. Methods: The aim of this study presents an analysis of final year results of UG degree students using data mining technique, which carried out in three of the private colleges in Tamil Nadu state of India. The primary objective of this research work is to apply the classification techniques to the prediction of the performance of students in end semester university examinations. Particularly, the decision tree algorithm C4.5 (J48), Bayesian classifiers, k Nearest Neighbor algorithm and two rule learner’s algorithms namely OneR and JRip are used for classifying the performance of students as well as to develop a model of student performance predictors. Results: The result of this study reveals that overall accuracy of the tested classifiers is above 60%. In addition classification accuracy for the different classes reveals that the predictions are worst for distinction class and fairly good for the first class. The JRip produces highest classification accuracy for the Distinction. Classification of the students based on the attributes reveals that prediction rates are not uniform among the classification algorithms. Also shows that selected data attributes have found to be influenced the classification process. The results showed to be satisfactory. Improvements: The study can be extended to draw the performance of other classification techniques on an expanded data set with more distinct attributes to get more accurate results.

Keywords

Classification Algorithm, Classifiers, Comparative Analysis, Educational Data Mining (EDM), Predicting Student Performance

Full Text

   

J. Manimaran ¹, T. Velmurugan ²

Affiliations
1 Bharathiar University, Coimbatore - 641 046, Tamil Nadu, IN
2 Research Department of Computer Science, D. G. Vaishnav College, Chennai - 600106, Tamil Nadu, IN

Abstract

Objective: Association rule mining is one of the data mining process for discovering frequent item set between transaction databases. The main objective of this research work is statistically analyses the quality rules in the apriori algorithm of association rule mining. Methods: An Interestingness measures is a subset of statistical method and it can give the solution for splitting interesting rules within huge association rules. Currently, it has shown around hundred and above measures. Specifically, this study is to concentrate on eight measures such as lift, chi-square, hyper-lift, hyper-confidence, conviction, coverage, leverage and cosine. In this analysis is performed in two places of real databases whereas Agriculture and Medical domain. Findings: At the experimental results, the proposed system is rectified that the problem many interesting rules are eliminated in satisfying the threshold value of support and confidence. Therefore, the user do not confirm that the strength of interest rules may be least by setting the low threshold value. The comparison and correlation measures also obtained along with the interesting rules. There are some measures outperformed than other and thus measures can mostly correlate with the order lift, chi-squared, hyper-lift, hyper-confidence and conviction. The performance of this work is consistently checking in difference size of transaction databases in addition to we identify the unresolved problem of apriori algorithm. Conclusion: Finally, this research concludes that statistical interestingness measures are really helpful for finding interesting rules among large association rules.

Keywords

Apriori Algorithm, Association Rule Mining, Interestingness Measures.

Full Text

   

S. Mahalakshmi ¹, T. Velmurugan ¹

Affiliations
1 PG and Research Department of Computer Science, D. G. Vaishnav College, Chennai- 600 106, Tamil Nadu, IN

Abstract

Background/Objectives: Image segmentation is one of the fundamental techniques in image processing. During past few years, the image processing mechanisms are extensively used in various medical fields for early stage detection, separation and identification of diseases; in this, the time consumption is an important criteria to discover the diseases for the patient. Methods/Statistical Analysis: This research work analyses about the detection and separation of brain tumor through Magnetic Resonance Imaging (MRI) medical images using Particle Swarm Optimization (PSO), a heuristic global optimization method based on swarm intelligence. The algorithm is widely used and rapidly developed for its ease implementation. This work has four stages that includes conversion, implementation, selection and extraction. Findings: The research work starts with converting the Digital Imaging and Communications in Medicine (DICOM) into image file format which is the first stage. Applying the PSO algorithm with the change in the values of n (segmentation level) is the second stage. Based on the time, selecting the best resultant images is the third stage. The final stage is extraction of tumor affected region with the suitable filtering techniques. The research work takes the axial and coronal plane of the Magnetic Resonance (MRI) images. Finally, this work concludes with the extraction of the resultant image, which is taken as input, and using the best filtering technique the affected region is easily separated and identified efficiently. The work also identifies the best suitable plane for the PSO algorithm. Applications/Improvements: The same PSO algorithm is applied to find the size and the type of calcifications in MRI brain images and are extracted in future. The possibilities of using other algorithms are also considered for further implementation.

Keywords

Brain Tumor, MRI Images, PSO Algorithm, Segmentation

Full Text

   

A. Naveen ¹, T. Velmurugan ¹

Affiliations
1 Department of Computer Science, D. G. Vaishnav College, Arumbakkam, Chennai – 600106, Tamil Nadu, IN

Abstract

Background/Objective: The role of clustering is significant to analyze different kind of applications of its techniques. Similar data are grouped into one and they formed as a cluster. Dissimilar data are grouped into another form in other cluster. Data clustering is an important and active research applied in many fields including multivariate analysis in statistics and some other areas like pattern recognition and machine learning etc. Methods/Statistical Analysis: Boundary detection and outlier analysis is an important concept for pre-processing the data. The boundary considers only pixels lying on and near edges and use of gradient or Laplacian to preliminary processing of images. To find the outlier in a group of patterns is a well-known problem in Data Mining (DM). An outlier is a pattern which is different with respect to the rest of the patterns in the data. The k-Means is one of the familiar clustering methods used by different researchers to find the well-formed clusters. Magnetic Resonance Imaging (MRI) uses a magnetic field and radio waves to create detailed images of the organs and tissues within human body. The k-Means algorithm is used to find the tumor by applying the boundary detection and outlier techniques in this research work in MRI brain images. Findings: The main goal of this research work is to extract the tumor (Calcification) in an MRI brain image by means of clustering pixels to fortify the quality of clustering algorithm. The results of the MRI brain images are analyzed and identified by the proposed algorithm. The result produced by simple k-Means algorithm is very useful to find the tumor in MRI images perfectly. Application/Improvements: The MRI brain images are analyzed and implemented by other methods like classification and some other techniques in future.

Keywords

Image Clustering, Image Preprocessing, k-Means Algorithm, MRI Imagery, Method

Full Text

   

E. Venkatesan ¹, T. Velmurugan ¹

Affiliations
1 PG and Research Department of Computer Science, D.G. Vaishnav College, Chennai-600106, Tamil Nadu, IN

Abstract

Backgrounds/Objectives: Data Mining (DM) techniques are extremely utilized for the extraction of useful information which is available in data warehouses and other database repositories. In medical diagnose, the role of DM approach rises quick recognition of disease over symptoms. To classify the medical data, a number of DM techniques are used by researchers. One of such techniques is classification. The classification algorithms predict the hidden information in the medical domain. The breast cancer is the very dangerous disease for women in developed countries like India. Most of the women death happens in the world, they are affected by the breast cancer. Methods/Statistical Analysis: The role of classification is importantin the real world applications in every field. Classification is used to classify the elements permitting to the features of the elements through the predefined set of classes. This research work analyses the breast cancer data using classification algorithms namely j48, Classification and Regression Trees (CART), Alternating Decision Tree (AD Tree) and Best First Tree (BF Tree). Findings: To find the performance of classification algorithms, this work uses cancer data as input. Particularly, this work is carried out to compare the four decision tree algorithms in the prediction of the performance accuracy in breast cancer data. All the algorithms are applied for breast cancer data to classify the data set for classification and prediction. Among these four methods, this work concludes the best algorithm for the chosen input data on decision tree supervised learning algorithms to predict the best classifier. Applications/Improvements: The breast cancer data is analyzed by taking the images using the same algorithms in future. Also, the microcalcifications of the breast cancer imagery are to be investigated in the same work.

Keywords

CART Algorithm, Classification Algorithms, Decision Trees, J48 Algorithm

Full Text

   

K. Saravananathan ¹, T. Velmurugan ²

Affiliations
1 SRM Arts and Science College, SRM Nagar, Kattankulathur - 603203, Tamil Nadu, IN
2 PG and Research Department of Computer Science, D.G. Vaishnav College, 833, E.V.R. Periyar High Road, Arumbakkam, Chennai - 600106, Tamil Nadu, IN

Abstract

Backgrounds/Objectives: Huge medical datasets available in various data repositories which are used for real world applications. To visualize the useful information stored in data warehouses, the Data Mining (DM) methods are enormously utilized. One of such domain is medical domain, in which the function of DM approach raises speedy recovery of sickness over indications. On the way to categorize and predict symptoms in medicinal data, a variety of DM methods are utilized by different researchers. From many techniques of DM, classification is one of the main techniques. The classification techniques classify the unseen information in all areas including medical diagnostic field. The very dangerous disease in medicinal field is diabetes disease which is affected for many peoples in popular countries like India. Methods/Statistical Analysis: The impact of categorization is very important in authentic earth applications in all fields. To categorize the rudiments allowing to the applications of the elements during the predefined set of modules are used by classification methods. Very popular classification algorithms J48, Support Vector Machines (SVM), Classification and Regression Tree CART and k-Nearest Neighbor (kNN) for diabetic data are used for this research work. Findings: To discover the presentation of these classification methods, diabetic data as an input. For the most part, this research work is supported out to associate the techniques in the calculation of the presentation accurateness in diabetic data. The above mentioned techniques are used for diabetic data to categorize its accuracy in terms of its performance. Methods: The conclusion of this research work is choosing the top algorithm for the input data for the best classifier. Applications/Improvements: Some of other algorithms are analyzed using the same data set for the similar type of results is discussed in future. Also, some of the clustering algorithms are applied using the same data set to find highly affected diabetic patients.

Keywords

CART Algorithm, Classification, J48 Algorithm, kNN Algorithm, SVM Algorithm.

Full Text

   

S. DeepaLakshmi ¹, T. Velmurugan ²

Affiliations
1 Bharathiar University, Coimbatore - 641046, Tamil Nadu, IN
2 PG and Research Department of Computer Science, D. G. Vaishnav College, Chennai - 600106, Tamil Nadu, IN

Abstract

Background/Objectives: Feature Selection is a process of selecting features that are relevant which is used in model construction by removing redundant, irrelevant and noisy data. A typical application of Text Mining is classification of messages and e-mails into spam and ham. Methods/Statistical Analysis: This article gives a comprehensive overview of the various Feature Selection methods for Text Mining. Various Filter methods like Pearson Correlation, Chi-square, Symmetrical Uncertainty and Mutual Information are applied to select the optimal set of features. Findings: Filter Feature Selection methods are used to classify Text data. Various Classification algorithms are applied using the optimal set of features obtained. The accuracy of classification algorithms are verified based on the chosen data set. Novelty/ Improvements: A comparative study of various filter methods for Feature Selection and classification algorithms for performance evaluation is conceded in this research work.

Keywords

Chi-Square, Feature Selection, Filter Method, Mutual Information, Pearson Correlation.

Full Text

   

D. Sumathi ¹, T. Velmurugan ², S. Nandakumar ², S. Renugadevi ²

Affiliations
1 School of Information Technology & Engineering, VIT University, Vellore, TamilNadu, IN
2 School of Electronics Engineering, VIT University, Vellore, TamilNadu, IN

Abstract

Objectives: Consolidating any two dissimilar networks leads to a formation of a heterogeneous wireless network. Vision to achieve distinct networks to get converged such that characterizing of the upcoming wireless networks becomes a reality. This in turn ushers in vertical handoff such that handoff among different technologies is efficient. And at the same time it is so smooth that naming it as seamless is justifiable. Pragmatic consideration of the network characteristics and the dynamics are very essential to choose one of the best available network handoff decisions. Methods: Stating conventionally, when a mobile client is roaming, a single criterion- for example, received signal strength is employed to realize the vertical handoff. But, single criterion consideration is not sufficient and taking into account of other parameters is needed for a proper handoff decision. The present paper puts forth a strong load balancing algorithm M-OPTF that is based on dynamic weight assignment technique for allocation of fresh calls and handoff calls. Findings: In the present scheme assignment of weight factor is done to attachment points basing on the distribution of load on them. In addition it contains a process for handoff decision that takes into account the velocity of mobile node to do triggering of handoff. The put forth M-OPTF algorithm is compared along with Remote Sensing Systems (RSS) and OPT-F algorithm in this paper. Application/Improvements: The simulation results confirm the fact that M-OPTF algorithm apart from balancing the network load it can effectively decrease congestion in network, reduce the number of handoffs that are unnecessary, and increases the battery life time, thus improve the overall performance of the system.

Keywords

Dynamic Weight Assignment, Load Balancing, Unnecessary Handoffs, Velocity; Vertical Handoff, Weight Factor.

Full Text

   

S. Nandakumar ¹, T. Velmurugan ¹, K. S. Preetha ¹, T. R. Suresh Kumar ¹

Affiliations
1 School of Electronics Engineering, VIT University, Vellore, TamilNadu, IN

Abstract

Background/Objectives: Next generation heterogeneous network will comprise multi access systems with multiple service support. Call management scheme is necessary to provide low call dropping probability for high priority services. Call management scheme for mixed cells (i.e., a 4G cell with embedded WLANs) increase cellular system capacity and reduce dropping probability of cellular-to-WLAN vertical handoff calls. Methods: We propose an efficient call management scheme which considers all possible vertical handoff scenarios and provides the maximum usage of WLAN. A blocked request in WLAN is taken back by the overlaying cellular system, if channels are available. Several existing models do not reflect the effect of configuration of neighbor cells which is important for cell planning for a cellular operator. Findings: Call management scheme is executed by diverting every request of WLAN that was blocked, to cellular system. Improvements/ Applications: The proposed call management scheme shows that new call and handoff call dropping probability is decreased when compare with Complementary-WLAN scheme (C-WLAN) and also increases in system throughput.

Keywords

Blocking and Dropping Probability, C-WLAN, Call Management Scheme, WLAN.

Full Text

   

E. Venkatesan ¹, T. Velmurugan ¹

Affiliations
1 PG and Research Department of Computer Science, D.G. Vaishnav College, Chennai-600106, IN

Abstract

Objectives/Backgrounds: The breast cancer has increased significantly in the last few years. It is one type of cancer and is the second deadliest disease in the world wide. Recently, cancer is diagnosed by various test such as mammography, ultrasound, etc. Mammography is used to breast imaging to help in detecting breast cancer. Methods/Statistical Analysis: The Mammogram images are taken for the analysis to find the tumor affected regions by data mining techniques in this research work. This work uses the Median filter method for noise removal and Gaussian filter for image enhancement of preprocessing the images. The k-Means algorithm, which is easily detected and extracts tumor area by means of intensity values by segmenting the mammography images. Two types of mammography images; normal and abnormal are given as input to the algorithms. After clustering the images by k-Means algorithm, the results found are classified by J48, JRIP, Support Vector Machines (SVM), Naïve Bayes and CART algorithms to verify the accuracy of the results based on its pixel values. Findings: The performance of taken classification algorithms is compared and find out the best classifier in terms of its accuracy, sensitivity and specificity. Improvements: In the future, the other classifiers and feature selection algorithms are applied to extract the mammography images. Also, it gives more than fifty images for analysis.

Keywords

Classification Accuracy, Classification Algorithms, k-Means Algorithm, Mammogram Images.

Full Text

   

M. P. Sukassini ¹, T. Velmurugan ²

Affiliations
1 Bharathiyar University, Coimbatore -641046, Tamil Nadu, IN
2 PG and Research Department of Computer Science, D. G. Vaishnav Chennai - 600106, Tamil Nadu, IN

Abstract

Background/Objectives: Images are the most effective way of revealinginformation to the world. Images are analysed and processed by computerized techniquesto extract hidden information available in it. Innumerable techniques are available to process the images. In numerous fields, the processed images are used for decision making. In medical field, automated detection and quantitative analysis of the radiological images and other images are processed by Computer Aided Diagnosis(CAD) tool to detect the abnormalities present in images. Methods/Statistical Analysis:Segmentation is one of the techniques used in CAD which play a vibrant role in processing the images. It is a process in which regions/ features sharing related characteristics are recognized or grouped together to interpret the images. Each segment divulges some information to the users. Segmentation techniques vary from image to image and areapplied to the images depending upon the problems being solved.This research work analyses about the segmentation techniques which are used in medical field. Findings: The results of the various research papers are discussed based on the approaches of hybrid algorithms applied in segmentation techniques. Also, this research work analyses about theuse of various segmentation techniques applied in mammogram images using CAD tool which assist the radiologist to interpret the suspicious regions and to know the impact of the diseases in patients for the suitable prediction. Application/Improvements: The segmentation techniquesare implemented to analyse mammogram images in future by means of a practical approach.

Keywords

Deformable Method, Hybrid Techniques, Image Segmentation, Mammogram Images

Full Text