Development of a software module for the identification of the emotional state of the user

Abstract

A huge number of spheres of human activity leads to the emergence of information re-sources that reflect social communication. The study of the identification of emotions in text communication is an actual direction of research in the field of natural language processing and machine learning. The main goal of the work is to develop a software module that implements algorithms and models that can automatically determine a person's emotional state based on text messages. This work is de-voted to the review of some models and an algorithm for improving data processing in the middle of text communication of users. One of the methods used in the work is the filtering method. The filtering method deter-mines the discussions of the text, which it records in the form of a hierarchical tree-like struc-ture. Discourse greatly simplifies the work and allows you to more accurately determine the emotion in the text. It also builds a semantic model, the data of which is obtained from the text communica-tion of users. Using the described structures, the filtering method finds emotional words re-corded in the database. The search is based on keywords. In turn, keywords are defined by case. The work deals with the issue of finding emotions in text messages and the development of a software module for its implementation. Two algorithms for determining emotions are considered - vector and Boolean. During the research, it was determined that the Boolean algorithm is most suitable for searching for emotional words. In the work, emotional words were found by identifying and analyzing the semantics of the sentence.