INTELLECTUAL SYSTEM DEVELOPMENT FOR USER SOCIALIZATION SUPPORT BY INTERESTS SIMILARITY

Abstract

The object of research is the process of socialization of individuals, because nowadays the task of socialization is very important and all modern social networks try to optimize and automate the socialization of various users using all popular modern technologies such as neural networks and user text analysis algorithms. The subject matter of the study is the methods and technologies for the search and formation of a list of relevant users by similarity of interests for socialization. Accordingly, the system user profile analysis is studied, namely the identification of the user by searching the human face in user photos using neural networks and analysing user information using fuzzy search algorithms and the Noisy Channel model. The goal of the work is to create an intelligent system for socialization of individuals based on fuzzy word search using the Noisy Channel model with algorithms for efficient distribution of textual information, and a convolutional neural network to identify users of the system. The following tasks were solved in the article: 1. Analyse modern and most well-known approaches, methods, tools and algorithms for solving problems of socialization of individuals by similar interests. 2. To development the general structure of a typical intellectual system of socialization of individuals by common interests. 3. To form functional requirements to the basic modules of structure of typical intellectual system of socialization of persons on common interests. 4. Develop an intelligent system of support for user socialization by similarity of interests based on neural networks, fuzzy search and Noisy Channel model and conduct experimental testing. The following methods are used: Levenstein's method; Noisy Channel model; N-gram algorithm; fuzzy search. The following results were obtained: the general structure of a typical intellectual system of socialization of individuals by common interests was built and described. The main purpose of the system is to create a new algorithm for analysing user information and finding the most suitable users, according to the analysed text based on existing algorithms such as Levenstein's algorithm, sampling algorithm, N-gram algorithm and Noisy Channel model. The template of asynchronous creation of a software product which will allow to create almost completely dynamic system also underwent further development. It is necessary to improve the convolutional neural network, which will allow efficient and dynamic search of human faces in the photo, and check the presence of existing people in the database of the system. Сonclusions: It was found that the implemented algorithm performs sampling approximately 10 times faster than the usual Levenstein algorithm. Also, the implemented in the system algorithm for forming a sample of users is more efficient and accurate by about 25-30% compared to the usual Levenstein algorithm.