MACHINE LEARNING SYSTEM BASED ON HYBRID MACHINE CHARACTER AND DEVELOPMENT METHOD THEREOF

摘要

The present invention relates to a machine learning system based on a hybrid machine character that a method for utilizing multiple personality of human is applied to machine learning and an implementation method thereof. The implementation method comprises the steps of: distinguishing one artificial intelligence into a plurality of machine characters to have classes and class characteristics different from each other based on learning data; applying test data to the plurality of distinguished machine characters to extract features for each state having a good classification success rate for each machine character; identifying a feature for each state consistent with a state of input data among the extracted features for each state when applying the input data to be recognized; and classifying classes of the input data by calling a machine character indicating a high recognition rate for the identified feature for each state of the machine characters.