Data Mining and Knowledge Engineering

Sanjay Garg ¹, Mahesh Panchal ²

Affiliations
1 Department of Computer Science & Engineering, Nirma Institute of Technology, Nirma University, Gujarat, IN
2 Department of Computer Engineering, Kalol Institute of Technology & Research Centre, Gujarat, IN

Abstract

High dimensional data contains large number of features (predictor attributes) compared to number of samples. As many of these features are irrelevant with class label, if any classification algorithm is directly applied on this dataset then model come out will be less accurate and will take much time for building, testing and applying on unseen data. Feature selection methods will select only those features which are relevant to class label. During feature selection procedure, set of features are generated and evaluated for its relevance with class. There are several methods proposed in literature for generation and evaluation of features. Each method has its own characteristic. In this paper experiment is carried out on three types of cancer gene expression datasets with different feature selection methods. Features are generated by ranker, heuristic and random search methods while they are evaluated by information gain, attreval and wrapper methods. A hybrid approach which combines ranker and subset based feature generation is also proposed. It shows that hybrid approach with wrapper evaluator gives best classification accuracy.

Keywords

Data Mining, Classification, Feature Selection, Wrapper Evaluators.

Full Text

     

Urmi Vasishnav ¹, Deven Agravat ², Sanjay Garg ³

Affiliations
1 Department of Information Technology, A D Patel Institute of Technology, New Vallabh Vidyanagar, Gujarat, IN
2 Department of Information Technology, G H Patel college of Engineering & Technology, New Vallabh Vidyanagar, Gujarat, IN
3 Department of Computer Engineering, Institute of Technology, Nirma University, Ahmedabad, Gujarat, IN

Abstract

With the rapid growth of World Wide Web and Internet, a problem of improving web server’s performance is having primary importance. When a web user requests any resource from web server, he/she experiences some amount of delay which is known as User Perceived Latency. Web caching is an effective and economical technique to reduce UPL. Performance of web server can still be improved by predicting a future request from the current page and then move it to the web server’s cache. This concept is known as web cache pre-fetching. Web cache pre-fetching using markov tree and multidimensional matrix have limitations regarding to time complexity and memory, respectively. In this paper, an attempt has been made to use markov model and concept of stationary distribution for web cache pre-fetching. A comparison has been presented for proposed algorithms with Least Recently Used (LRU). Empirical results based on implementation confirm that proposed algorithms are better than LRU.

Keywords

Markov Model, User Perceived Latency, Web Cache Pre-Fetching.

Full Text