Начальные этапы проектирования системы сбора и предиктивного анализа данных социальных медиа

Abstract

This article discusses the design of a system for collecting and predictive analysis of social media. With the development of the Internet, as well as social media, it has become easier to access and distribute information because network users themselves are both creators and recipients of diverse information. To gain new knowledge that can be useful to users of social media, it is possible to use predictive analytics – a set of statistical analysis methods that extract new information from current and historical data. This method of analyzing social media data is at the stage of its development. Predictive analytics is based on automatic search for connections, anomalies and patterns between various factors. To form a predictive model, a large set of statistical modeling methods, data mining, machine learning, neural networks and other mechanisms are used. Together with various methods of collecting information from Internet resources, such as parsing and social network APIs, predictive analytics can offer the most interesting sources of information for the user. In order to combine the methods of predictive analysis and data collection methods, it is necessary to take a detailed approach to the system design process. The paper proposes a formal description of the data that a future system uses. In addition, the general architecture and algorithm of functioning are highlighted. Special attention is paid to a detailed description of one of the main parts of the system (the collection subsystem). The obtained results will be used in further design, and it is planned to further study the analytics subsystem. Subsequent work on this topic will make it possible to detail the architecture and algorithm of functioning.