The Use of Entropy Measure for Higher Quality Machine Learning Algorithms in Text Data Processing

A large number of text data are regularly published in social networks and the media. Processing and analysis of such information is an highly required direction. This paper focuses on the way to use the entropy measure when dealing with big volumes of text data in classification. The used entropy measure stands for algorithm quality criteria when defining a class in a set of data. The work also features a comparative efficiency analysis of the proposed approach used with different number of classes and diverse machine learning algorithms. The Entropy measure can be used to build a committee of voting algorithms afterwards as well.