Approaches to the solution to the problem of news-based events forecasting

Abstract

An overview of the areas of application of approaches and methods of forecasting events based on past events. The substantiation of urgency of a theme is given and possibilities concerning application of results of work are resulted. Requirements for incoming news regarding their quality are defined. It is noted that there are four key criteria for the quality of the media, which are often two-component, namely: the relevance of news, providing the context in which the event, compliance with professional standards and a variety of materials. The key stages of working with data in order to obtain knowledge from them for forecasting events are identified. These include pre-processing of data (reduction to a standardized view that will understand and be able to process the algorithm), their analysis and the forecasting process itself. The spheres of application of associative series and Markov processes for search of causal relations, and time series for definition of the period of occurrence of an event with the set probability are specified. Ref. 7, pic. 2.

Keywords

• event forecasting
• data mining
• news quality
• computer linguistics
• time series
• associative rules