Voting based learning classifier system for multi-label classification
Kaveh Ahmadi Abhari, Ali Hamzeh, Sattar Hashemi · 2011
Learning Classifier Systems (LCSs) are rule-based systems with a discovery mechanism to find additional meaningful rules according to the results of its previous experiments. LCSs were designed to deal with both single and multistep problems. In the first category, almost all major studies focus on the single-label classification problems. However, there are more complex problems that require multi-label classification. The aim of this study is to take advantage of the power and ability of LCSs for solving multi-label classification problems. The main idea behind this research is to guide the discovery mechanism by a prior knowledge. This prior knowledge is defined as a voting mechanism that realizes the quality of the existing rules and is used in discovering new rules. Our proposed system is called Voting Based LCS (VLCS). The experimental results show the proposed method has potential for future research and progress.