期刊名称：International Journal of Advanced Research in Computer Engineering & Technology (IJARCET)
印刷版ISSN：2278-1323
出版年度：2016
卷号：5
期号：8
页码：2294-2301
出版社：Shri Pannalal Research Institute of Technolgy
摘要：Generally, decision Making and Modeling in high pressure, fast paced, complicated environments iscommonly confounded by the lack of the choice model to capture the requisite style of true in a very ungenerousmanner. This analysis focuses on a sophisticated belief rule-based decision model and proposes a dynamic ruleactivation (DRA) technique to deal with each problem at the same time. DRA relies on “smart” rule activation,wherever the active optimum rules are selected in a very dynamic way to seek for a balance between the wholenessand inconsistency within the rule-base generated from sample knowledge to attain a stronger performance. A seriesof case studies demonstrate however the utilization of DRA improves the accuracy of this advanced rule-based callmodel, while not compromising its potency, particularly once handling multi-class classification datasets. DRA hasbeen tested to be beneficial to pick out the foremost appropriate rules or knowledge instances rather than aggregatinga complete rule-base. Beside the work performed in rule-based systems, DRA alone are often considered a genericdynamic similarity method that may be applied in several domains. During this analysis Continuous ant Colonyoptimization (CACO) is applied within the context of Extended Belief Rule-Bases (E-brbs). The experimentalresults illustrated during this analysis demonstrate however the utilization of DRA will improve the accuracy of EBRBprimarily based decision support models.
关键词：Rule-based processing; incompleteness; inconsistency; decision support; Continuous Ant Colony;Optimization
