Development of Judgment Classification Models using Machine Learning

Abstract

With the world becoming more and more reliant on technology, we are transitioning from a society that values rational evaluation over intuitive thinking to one in which both of those methods coexist. AI devices rely solely on rational evaluation and machine learning allows us to focus on intuition. The task of intelligence is to deduce which method should be relied upon when solving various problems via the establishment of realistic judgments, according to what kind it identifies as being best for that particular problem. However, human judgments cannot simply be quantitatively compared and ranked by a computer according to conditions set by algorithms because certain difficult-to-measure criteria are not easily passable through algorithm systems such as ethics and common sense. In this research, the authors focus on developing judgment classification models using random forest and support vector machine. The authors attempt to test the effectiveness of sentiment proportions as features in judgment classification models.