Medical Pattern Classification Using a Novel Binary Similarity Approach based on an Associative Classifier

Osvaldo Velázquez-González¹Antonio Alarcón-Paredes¹Cornelio Yáñez-Márquez^1,2*

• ¹Centro de Investigación en Computación, Instituto Politécnico Nacional (CIC-IPN), Mexico City, México, Mexico
• ²National Polytechnic Institute (IPN), Mexico City, Mexico

Classification is a central task in machine learning, underpinning applications in domains such as finance, medicine, engineering, information technology and biology. However, machine learning pattern classification can become a complicated or totally inexplicable task for current robust models due to the complexity that objective datasets can have, which is why there is a high interest in achieving a good performance classification. On the other hand, there is a need in particular cases to achieve such performance but maintaining a certain explainability in the operation and decisions of classification algorithms, which can become complex to have. For this reason, an algorithm is proposed that is robust but simple and highly explainable in datasets mainly related to medicine and with complexity in class imbalance. The main contribution of this research is a novel approach to a new machine learning classification algorithm based on binary string similarity, which is competitive but simple, interpretable and transparent, since it is known why a pattern is classified within a certain class. Therefore, a comparative study of the performance of the best-known state-of-the-art classification algorithms and the proposed model is presented. The experimental results demonstrate the benefits of the proposal of this research work, which were validated with the purpose of finding significant differences in performance using statistical hypothesis tests.