Microarray data with reference to gene expression profiles have provided some valuable results re... more Microarray data with reference to gene expression profiles have provided some valuable results related to a variety of problems, and contributed to advances in clinical medicine. Microarray data characteristically have a high dimension and small sample size, which makes it difficult for a general classification method to obtain correct data for classification. However, not every gene is potentially relevant for distinguishing the sample class. Thus, in order to analyze gene expression profiles correctly, feature (gene) selection is crucial for the classification process, and an effective gene extraction method is necessary for eliminating irrelevant genes and decreasing the classification error rate. The purpose of gene expression analysis is to discriminate between classes of samples, and to predict the relative importance of each gene for sample classification. In this paper , irrelevant genes are eliminated in two stages, employing correlation-based feature selection (CFS)as an evaluator and binary particle swarm optimization (BPSO)as a search technique at the first phase and in the second phase of elimination it (CFS+PSO) combines Quick-Reduct algorithm and forms an integrated filter method. Since the data consist of a large number of redundant features, an initial redundancy reduction of the gene is done to enable faster convergence. Then Rough set theory is employed to generate reducts, which represent the minimal sets of non-redundant gene capable of discerning between all objects, in a multi-objective framework. The effectiveness of the proposed approach was verified on two different multi-class microarray datasets using K-nearest neighbour (K-NN), Naïve Bayes( NB) and J48 classifiers with leave-one-out cross-validation (LOOCV) method.
The data mining technology is used to identifying patterns and information from a huge quantity o... more The data mining technology is used to identifying patterns and information from a huge quantity of data. In a single repository data base where data is stored in central site, then applying data mining algorithms on these data base, patterns are extracted, which is clearly implausible and untenable for many realistic problems and databases. To deal with these complex systems has revealed opportunities to improve distributed data mining systems in a number of ways. Furthermore, in certain situations, data may be inherently distributed and cannot be merged into a single database for a variety of reasons including security, fault tolerance, legal constraints, competitive reasons, etc. In such cases, it may not be possible to examine all of the data at a central processing site to compute a single global model. Here, we develop techniques that scale up to large and possibly physically distributed databases. Meta-learning (learning from learned knowledge) -a technique dealing with the problem of computing a global classifier from large and inherently distributed databases. This paper, describes meta-learning and JAM system (Java Agents for Metalearning), which is an agent-based meta-learning system for large-scale data mining applications. Several important desiderata of data mining systems are addressed (i.e., scalability, efficiency, portability, compatibility, adaptivity, extensibility and effectiveness) and a combination of AI-based methods and distributed systems techniques are presented. We applied JAM on the real-world data mining task of modeling and detecting credit card fraud with notable success.
Microarray data with reference to gene expression profiles have provided some valuable results re... more Microarray data with reference to gene expression profiles have provided some valuable results related to a variety of problems, and contributed to advances in clinical medicine. Microarray data characteristically have a high dimension and small sample size, which makes it difficult for a general classification method to obtain correct data for classification. However, not every gene is potentially relevant for distinguishing the sample class. Thus, in order to analyze gene expression profiles correctly, feature (gene) selection is crucial for the classification process, and an effective gene extraction method is necessary for eliminating irrelevant genes and decreasing the classification error rate. The purpose of gene expression analysis is to discriminate between classes of samples, and to predict the relative importance of each gene for sample classification. In this paper , irrelevant genes are eliminated in two stages, employing correlation-based feature selection (CFS)as an evaluator and binary particle swarm optimization (BPSO)as a search technique at the first phase and in the second phase of elimination it (CFS+PSO) combines Quick-Reduct algorithm and forms an integrated filter method. Since the data consist of a large number of redundant features, an initial redundancy reduction of the gene is done to enable faster convergence. Then Rough set theory is employed to generate reducts, which represent the minimal sets of non-redundant gene capable of discerning between all objects, in a multi-objective framework. The effectiveness of the proposed approach was verified on two different multi-class microarray datasets using K-nearest neighbour (K-NN), Naïve Bayes( NB) and J48 classifiers with leave-one-out cross-validation (LOOCV) method.
The data mining technology is used to identifying patterns and information from a huge quantity o... more The data mining technology is used to identifying patterns and information from a huge quantity of data. In a single repository data base where data is stored in central site, then applying data mining algorithms on these data base, patterns are extracted, which is clearly implausible and untenable for many realistic problems and databases. To deal with these complex systems has revealed opportunities to improve distributed data mining systems in a number of ways. Furthermore, in certain situations, data may be inherently distributed and cannot be merged into a single database for a variety of reasons including security, fault tolerance, legal constraints, competitive reasons, etc. In such cases, it may not be possible to examine all of the data at a central processing site to compute a single global model. Here, we develop techniques that scale up to large and possibly physically distributed databases. Meta-learning (learning from learned knowledge) -a technique dealing with the problem of computing a global classifier from large and inherently distributed databases. This paper, describes meta-learning and JAM system (Java Agents for Metalearning), which is an agent-based meta-learning system for large-scale data mining applications. Several important desiderata of data mining systems are addressed (i.e., scalability, efficiency, portability, compatibility, adaptivity, extensibility and effectiveness) and a combination of AI-based methods and distributed systems techniques are presented. We applied JAM on the real-world data mining task of modeling and detecting credit card fraud with notable success.
Uploads
Papers by Dr. Sujata Das