Increasing the Efficiency of IBLR_ML Algorithm by Using Multi-Agent Model

Multi-label classification is an extension of conventional classification in which each instance is assumed to belong to exactly one among a finite set of candidate classes. Multi-label text categorization problem is the prime motivation of multi-label classification, where each document may belong to several predefined topics simultaneously. In multi-label learning, the training set is composed of instances that each associated with a set of labels, and the task is to predict the label sets of unseen instances through analyzing training instances with known label sets. A novel method of Multi-label classification is combining instance-based learning and logistic regression for multi-label classification . This algorithm suffers from high computational complexity. In this paper, multi-agent model is used for this algorithm to access more efficiency specially when the training set is extensive or the number of label or attributes is many. Multi-Agent Systems utilizes parallel techniques, and decrease considerably the consumption time of the algorithm. This method is experienced on five different data set and the results have been compared to sequential method. The results signifying the increase of almost 2-times speed for multi agent system.

Keywords:

Multi-label classification , multi-agent programming , parallel technique