ANALYSIS OF THE EFFECTIVENESS OF APPROACHES TO FILTERING SOCIAL MEDIA POSTS ON THE TOPIC OF EMIGRATION

Goryachkin Boris Sergeevich, Chirkin Konstantin Nikolaevich

Problem statement. The problem of emigration of the population from the Russian Federation is becoming more and more relevant, and in order to take measures to reduce it, it is necessary to understand the factors of emigration of the population. The authors propose an analysis of the texts of social network posts to identify factors of population emigration. However, since posts are collected using keywords that can be semantically rich, a large number of irrelevant posts may be included in the sample for analysis. In this regard, it is necessary to decide on the most effective approach to filtering social network posts, since, on the one hand, it is necessary to cut off irrelevant posts as much as possible to reduce analysis time, and on the other hand, to include the maximum number of relevant posts in the sample so as not to miss out on valuable information. Target. Determining the most accurate approach to filtering social network posts on the topic of emigration. Results. An algorithm based on an automated structural approach and a neural network approach were implemented. The accuracy of filtering posts using each approach was measured. A comparative analysis of post filtering approaches was carried out. Practical significance. The most effective approach to filtering posts will be used in a technological algorithm for identifying emigration factors based on the analysis of texts of social network posts and will increase the effectiveness of this algorithm. The algorithm, in turn, will allow public authorities, as well as researchers, to identify factors of emigration of the population in order to develop measures to reduce it. Such measures should be more effective, as they will be developed taking into account public opinion.

Keywords:Emigration factors, neural network approach, neural network, automated structural approach, social networks