摘要
Over the past few decades, several covering algorithms have been developed to automate the acquisition of knowledge from a set of examples. These algorithms employ a specific search process for extracting IF-THEN rules. They rely heavily on statistical measures to guide the search for rules; however, the information carried by these measures is limited and does not always lead to the best results. This paper presents two new algorithms which employ a new knowledge representation scheme to optimize the search and reduce the role of statistical measures, namely RULES-5 Plus for classification problems (discrete outputs) and RULES-F Plus for control applications (numerical outputs).
- 出版日期2011-11