Bikash Kanti Sarkar

A genetic algorithm-based rule extraction system

Individual classifiers predict unknown objects. Although, these are usually domain specific, and lack the property of scaling up prediction while handling data sets with huge size and high-dimensionality or imbalance class distribution. This article introduces an accuracy-based learning system called DTGA (decision tree and genetic algorithm) that aims to improve prediction accuracy over any classification problem irrespective to domain, size, dimensionality and class distribution. More specifically, the proposed system consists of two rule inducing phases. In the first phase, a base classifier, C4.5 (a decision tree based rule inducer) is used to produce rules from training data set, whereas GA (genetic algorithm) in the next phase refines them with the aim to provide more accurate and high-performance rules for prediction. The system has been compared with competent non-GA based systems: neural network, Naive Bayes, rule-based classifier using rough set theory and C4.5 (i.e., the base classifier of DTGA), on a number of benchmark datasets collected from UCI (University of California at Irvine) machine learning repository. Empirical results demonstrate that the proposed hybrid approach provides marked improvement in a number of cases.

Accuracy-based learning classification system

In order to implement a multi-category classification system, an efficient rule set is imperative for its investigation. In this paper, such a system is being introduced. In the first phase of its kind, the C4.5 rule induction algorithm is adopted to obtain useful rule set from classification problem, following a new data set partitioning approach. Next, the presented genetic algorithm (GA) is implemented to refine the learned rules in more efficient way. The resultant system has been compared with UCS (GA-based classification system) and C4.5 (non GA-based rule induction algorithm) on a number of benchmark data sets collected from UCI (University of California at Irvine) machine learning repository. Results demonstrate that the proposed genetic approach provides marked improvement in a number of cases.

A hybrid approach to design efficient learning classifiers

Recently, use of a Learning Classifier System (LCS) has become promising method for performing classification tasks and data mining. For the task of classification, the quality of the rule set is usually evaluated as a whole rather than evaluating the quality of a single rule. The present investigation proposes a hybrid of the C4.5 rule induction algorithm and a GA (Genetic Algorithm) approach to extract an accuracy based rule set. At the initial stage, C4.5 is applied to a classification problem to generate a rule set. Then, the GA is used to refine the rules learned. Using eight well-known data sets, it has been shown that the present work, in comparison to C4.5 alone and UCS, provides a marked improvement in a number of cases.

Two-warehouse inventory model on pricing decision

Abstract This paper develops a structural Economic Order Quantity (EOQ) model that evaluates the impact of a reduction rate in the selling price when the capacity of an Own Warehouse (OW) is limited and a Rented Warehouse (RW) is used, if needed. In many practical situations, there exists many factors to consider such as temporary price discounts, the limited capacity of OW houses, and transportation problems. Retailers need to build a new warehouse or rent other warehouses. However, from an economical point of view, they usually choose to rent other warehouses. However, in industries such as the textile and footwear industry or the food processing industry, after an certain time, the items decay or deteriorate. When this happens, a good manager wants to clear the stock by selling the huge items at reduced prices. In this paper the demand rate up to the start of deterioration is considered to be fixed and after that time, the new demand rate is considered to be a nonlinear increasing power function of the reduction rate. Possible associated profit functions affected by the trade off in inventory such as the own warehouse and rented warehouse holding costs, set up costs, the purchasing costs, the selling prices are discussed analytically and also solved using a Genetic Algorithm (GA) with numerical illustrations.

A conceptual distributed framework for improved and secured healthcare system

ABSTRACT In the twenty-first century, healthcare information technology has created the ability to electronically store, maintain, and move data across the world in a matter of seconds. It allows each healthcare provider to possess its own database of patient’s electronic medical record (EMR). The concept of ‘big data’ is now treated from different points of view covering its implications in many fields, including the healthcare sector. To achieve the wealth of health information, integrating, sharing, and availing data are the essential tasks that ultimately demand the concept of a distributed system. However, privacy and security of data are the matter of concerns, as data need to be accessed from various locations and users in the distributed system. The purposes of the present study are to provide an overview of big data and the effectiveness of big data in healthcare and to introduce a secured distributed e-healthcare system to provide better and secured health services across the nation.

A case study on partitioning data for classification

Designing accurate model for classification problem is a real concern in context of machine learning. The various factors such as inclusion of excellent samples in the training set, the number of samples as well as the proportion of each class type in the set (that would be sufficient for designing model) play important roles in this purpose. In this article, an investigation is introduced to address the question of what proportion of the samples should be devoted to the training set for developing a better classification model. The experimental results on several datasets, using C4.5 classifier, shows that any equidistributed data partitioning in between (20%, 80%) and (30%, 70%) may be considered as the best sample partition to build classification model irrespective to domain, size and class imbalanced.

A combined approach to tackle imbalanced data sets

Learning with imbalanced data causes high error-rates. Several approaches have been developed for addressing this problem. In this paper, a new learning model, integrating the C4.5 classifier and evolutionary algorithms, is introduced. To strengthen the model, two separate partitioning data sets are chosen for each original data set, by applying two distinct partitioning schemes proposed in this investigation, and these are used in sequence by the learning model. More specifically, the hybrid system first applies the base method C4.5 to produce a set of rules R from a training set say T_1, as constructed by the first data partitioning scheme. The R is then passed to the Genetic Algorithm to discover another set of rules say R_{GA} from another disjoint training set say T_2. T_2 is decided by the proposed second partitioning method. Finally, some informative rules of R_{GA} are included into R. The presented system is tested on several real data sets collected from the UCI machine learning repository and compared with standard C4.5. Experimental results show the good suitability of the system on imbalanced data sets. However, the model does not show negative performance on balanced data sets too.

Parallel Algorithm for LaGrange’s Interpolation on BSN-Mesh

The efficient mapping of numerical problems over parallel architectures is a desirable and challenging task. Bi-Swapped Networks (BSN) is a new class of network architecture related to the family of Swapped or Optical Transpose Interconnection System (OTIS) Networks. The architectural advancement of BSN ensures vertex symmetry and thus improves algorithmic efficiency. Further, it ensures regularity, modularity, scalability, robustness of the network. In the present article, parallel implementation of Lagrange’s interpolation over BSN-Mesh is performed. The architectural properties of BSN make it a desirable network for the efficient parallel mapping over it. The parallel algorithm for LaGrange’s interpolation is designed over BSN-Mesh and then analyzed accordingly. The proposed parallel algorithm claims to map LaGrange’s interpolation over BSN-Mesh in 12(n−1) electronic and five optical moves.

Big Data and Healthcare Data: A Survey

Big data and its analytics yield a lot of opportunities to make great progresses in many fields, ranging from economic and business activities to public administration, from national security to scientific researches and so on. However, the most noticeable point is that healthcare data has been recently identified as a prime example of big data. Undoubtedly, efficient use of healthcare resources has become a key factor in improving overall healthcare system. But for managing healthcare data and obtaining potential results, we need integration and sharing of data that ultimately demand the concept of distributed system. The paper in its first phase gives an overview on big data and healthcare data from different aspects. A review on the state-of-the-art distributed file system (Hadoop) is conducted in this stage too. The primary aim of this phase is to provide an overall picture on big data as well as healthcare data for non-expert readers. In the next phase, a cloud-based e-health system is proposed for the expert audiences. The expected promising characteristics as well as the managerial implications of the model are highlighted in the analysis section.

Classification system using parallel genetic algorithm

Classification task aims at predicting the value of the class attribute of new input data on the basis of a set of pre-classified samples. Traditional machine learning algorithms for classification are usually domain specific or produce unsatisfactory results whenever applied to classification problems with larger size or imbalanced data. Thus, to accumulate genuine useful knowledge for making decision, we introduce here a new intelligent knowledge discovery model, combining C4.5 (a decision tree-based rule inductive algorithm) with a new parallel genetic algorithm (GA) based on the idea of massive parallelism (MP). The model is named as CGAMP (C4.5 and GA based on MP). More specifically, the suggested model receives a base method C4.5 to produce rules which are then refined by the proposed parallel GA to provide more accurate rules. The strength of the developed system has been compared with pure C4.5 and a hybrid system (combining C4.5 and sequential genetic algorithm) on six real world benchmark data sets (collected from University of California at Irvine machine learning repository). The experimental results validate the effectiveness of the new model.