P. Deepa Shenoy

Dynamic Association Rule Mining using Genetic Algorithms

A large volume of transaction data is generated everyday in a number of applications. These dynamic data sets have immense potential for reflecting changes in customer behaviour patterns. One of the strategies of data mining is association rule discovery which correlates the occurrence of certain attributes in the database leading to the identification of large data itemsets. This paper seeks to generate large itemsets in a dynamic transaction database using the principles of Genetic Algorithms. Intra Transactions, Inter Transactions and Distributed Transactions are considered for mining Association Rules. Further, we analyze the time complexities of single scan technique DMARG (Dynamic Mining of Association Rules using Genetic Algorithms), with Fast UPdate (FUP) algorithm for intra transactions and E-Apriori for inter transactions. Our study shows that the algorithm DMARG outperforms both FUP and E-Apriori in terms of execution time and scalability, without compromising the quality or completeness of rules generated.

Aspect term extraction for sentiment analysis in large movie reviews using Gini Index feature selection method and SVM classifier

With the rapid development of the World Wide Web, electronic word-of-mouth interaction has made consumers active participants. Nowadays, a large number of reviews posted by the consumers on the Web provide valuable information to other consumers. Such information is highly essential for decision making and hence popular among the internet users. This information is very valuable not only for prospective consumers to make decisions but also for businesses in predicting the success and sustainability. In this paper, a Gini Index based feature selection method with Support Vector Machine (SVM) classifier is proposed for sentiment classification for large movie review data set. The results show that our Gini Index method has better classification performance in terms of reduced error rate and accuracy.

Evolutionary approach for mining association rules on dynamic databases

A large volume of transaction data is generated everyday in a number of applications. These dynamic data sets have immense potential for reflecting changes in customer behaviour patterns. One of the strategies of data mining is association rule discovery, which correlates the occurrence of certain attributes in the database leading to the identification of large data itemsets. This paper seeks to generate large itemsets in a dynamic transaction database using the principles of Genetic Algorithms. Intra transactions, Inter transactions and distributed transactions are considered for mining association rules. Further, we analyze the time complexities of single scan DMARG(Dynamic Mining of Association Rules using Genetic Algorithms), with Fast UPdate (FUP) algorithm for intra transactions and E-Apriori for inter transactions. Our study shows that DMARG outperforms both FUP and E-Apriori in terms of execution time and scalability, without compromising the quality or completeness of rules generated. The problem of mining association rules in the distributed environment is explored in DDMARG(Distributed and Dynamic Mining of Association Rules using Genetic Algorithms).

A Data Mining Approach for Data Generation and Analysis for Digital Forensic Application

With the rapid advancements in information and communication technology in the world, crimes committed are becoming technically intensive. When crimes committed use digital devices, forensic examiners have to adopt practical frameworks and methods to recover data for analysis which can pose as evidence. Data Generation, Data Warehousing and Data Mining, are the three essential features involved in the investigation process. This paper proposes a unique way of generating, storing and analyzing data, retrieved from digital devices which pose as evidence in forensic analysis. A statistical approach is used in validating the reliability of the pre-processed data. This work proposes a practical framework for digital forensics on flash drives.

Heartbeat biometrics for human authentication

Automated security is one of the major concerns of modern times. Secure and reliable authentication systems are in great demand. A biometric trait like the electrocardiogram (ECG) of a person is unique and secure. In this paper, we propose an authentication technique based on Radon transform. Here, ECG wave is considered as an image and Radon transform is applied on this image. Standardized Euclidean distance is applied on the Radon image to get a feature vector. Correlation coefficient between such two feature vectors is computed to authenticate a person. False Acceptance Ratio of the proposed system is found to be 2.19% and False Rejection Ratio is 0.128%. We have developed two more approaches based on statistical features of an ECG wave as our ground work. The result of proposed technique is compared with these two approaches and also with other state-of-the-art alternatives.

Classification of Neurodegenerative Disorders Based on Major Risk Factors Employing Machine Learning Techniques

Medical data mining has great potential for exploring the hidden patterns in the data sets of the medical domain. These patterns can be utilized for the classification of various diseases. Data mining technology provides a user-oriented approach to novel and hidden patterns in the data. The present study consisted of records of 746 patients collected from ADRC, ISTAART, USA. Around eight hundred and ninety patients were recruited to ADRC and diagnosed for Alzheimers disease (65%), vascular dementia (38%) and Parkinsons disease (40%), according to the established criteria. In our study we concentrated particularly on the major risk factors which are responsible for Alzheimers disease, vascular dementia and Parkinsons disease. This paper proposes a new model for the classification of Alzheimers disease, vascular disease and Parkinsons disease by considering the most influencing risk factors. The main focus was on the selection of most influencing risk factors for both AD and PD using various attribute evaluation scheme with ranker search method. Different models for the classification of AD, VD and PD using various classification techniques such as Neural Networks (NN) and Machine Learning (ML) methods were also developed. It is observed that increase in the vascular risk factors increases the risk of Alzheimers disease. It was found that some specific genetic factors, diabetes, age and smoking were the strongest risk factors for Alzheimers disease. Similarly, for the classification of Parkinsons disease, the risk factors such as stroke, diabetes, genes and age were the vital factors.

Classification of email using BeaKS: Behavior and keyword stemming

Spam mails are one of the greatest challenges faced by internet service providers, organizations and internet users in unison. Spam mails may be targeted, with a malicious intent or just as a commercial marketing activity - on the whole unwanted by everyone except the dispatcher. Spam filters continuously evolve as spammers go techno-savvy and creative. Machine learning algorithms have been popularly used for classifying and predicting mails as spam or ham (the good emails). This work presents a spam filter, BeaKS, with a focused preprocessing phase that weaves both the content of the email and two behavioral characteristics extracted from the email, to predict the category a mail belongs to: spam or ham. The accuracy of the proposed prediction model using Random Forests as the classifier is shown to be superior over other recent techniques. This approach is simple, easy to implement and reliable.

A Novel Data Generation Approach for Digital Forensic Application in Data Mining

With the rapid advancements in information and communication technology in the world, crimes committed are also becoming technically intensive. When crimes committed use digital devices, forensic examiners have to adopt practical frameworks and methods for recovering data for analysis as evidence. Data Generation, Data Warehousing and Data Mining, are the three essential features involved in this process. This paper proposes a unique way of generating, storing and analyzing data, retrieved from digital devices which pose as evidence in forensic analysis. A statistical approach is used in validating the reliability of the pre-processed data. This work proposes a practical framework for digital forensics on flash drives.

Authentication using Finger Knuckle Prints

Automated security is one of the major concerns of modern times. Secure and reliable authentication systems are in great demand. A biometric trait like the finger knuckle print (FKP) of a person is unique and secure. Finger knuckle print is a novel biometric trait and is not explored much for real-time implementation. In this paper, three different algorithms have been proposed based on this trait. The first approach uses Radon transform for feature extraction. Two levels of security are provided here and are based on eigenvalues and the peak points of the Radon graph. In the second approach, Gabor wavelet transform is used for extracting the features. Again, two levels of security are provided based on magnitude values of Gabor wavelet and the peak points of Gabor wavelet graph. The third approach is intended to authenticate a person even if there is a damage in finger knuckle position due to injury. The FKP image is divided into modules and module-wise feature matching is done for authentication. Performance of these algorithms was found to be much better than very few existing works. Moreover, the algorithms are designed so as to implement in real-time system with minimal changes.

Statistical Analysis for Human Authentication Using ECG Waves

Automated security is one of the major concerns of modern times. Secure and reliable authentication of a person is in great demand. A biometric trait like the electrocardiogram (ECG) of a person is unique and secure. In this paper we propose an authentication system based on ECG by using statistical features like mean and variance of ECG waves. Statistical tests like Z −test, t −test and χ 2 −tests are used for checking the authenticity of an individual. Then confusion matrix is generated to find False Acceptance Ratio (FAR) and False Rejection Ratio (FRR). This methodology of authentication is tested on data set of 200 waves prepared from ECG samples of 40 individuals taken from Physionet QT Database. The proposed authentication system is found to have FAR of about 2.56% and FRR of about 0.13%. The overall accuracy of the system is found to be 99.81%.