An Empirical Study on Classification Using Modified Teaching Learning Based Optimization

In this paper the modification to ‘Teaching–Learning BasedOptimization (TLBO) called Modified Teaching–LearningBased Optimization (MTLBO) based on particle swarmoptimization principle has been proposed. Unlike TLBO, thispopulation based method works on the effect of influence of ateacher on learners to find the optimum solution. The process ofMTLBO is divided into two parts, the first part consists of the‘Teacher Phase’ means learning from the teacher and the secondpart consists of the ‘Learner Phase’ means learner learns byinteracting with other learner having better knowledge and fromthe best learner knowledge treated as team leader among alllearners. The effectiveness of the method is tested on manybenchmark problems with different characteristics and theresults are compared with other population based methods andfinally it is implemented on classification using neural networkin data mining.