Biparti Majority Learning with Tensors

Abstract

In addition to the mislabeled training data that could interfere the effectiveness of learning, in a dynamic environment where the majority pattern changes, it is also difficult to perform training. We propose an efficient bipartite majority learning algorithm (BML) for categorical data classification with tensors on a single hidden layer feedforward neural network (SLFN). We adopt the resistant learning approach to avoid significant impact from data anomalies and iteratively conduct bipartite classification for majorities afterward. The bipartite algorithm can reduce the training time significantly while keeping competitive accuracy compared to previous resistant learning algorithms.