Fitness Function in Genetic Algorithm based Information Filtering - A Survey?

Information filtering systems have a vital role in exploiting the flurry of information that keeps on growing exponentially. This study focuses on improving the performance of genetic algorithms based information filtering. In GA, the fitness function is the performance measure for relevance judgement. Being the backbone of the evaluation process the selection of fitness function is vital in the design of a GA. The secondary data gathered from past studies lead to the conclusion that Cosine Coefficient is the better similarity measure among other alternatives like Euclidean Distance, Dice Coefficient, Jaccard Coefficient and Inner Product.