Breast cancer detection using a multi-objective binary Krill Herd algorithm

In this paper, an algorithm is presented for extracting fuzzy rules from the Breast Cancer dataset. To extract fuzzy rules, an imitation based evolutionary algorithm is used called Krill Herd (KH). The KH algorithm is converted to a binary algorithm here, and is used for the classification problem with innovation, named Binary Krill Herd-based Fuzzy Rule Miner (BKH-FRM). Choosing the best krill and local best of the Krills in each generation are performed according to a new multi-objective function. This algorithm achieves a higher accuracy than others with few rules and little sum of the rules lengths.