Analysis of the Impact of News on Stock Market Quotes

The main purpose of the study, described in the paper, is to determine the impact of news on stock market quotes. To perform the purpose, the Yandex Company, its shares, circulating on the Moscow exchange (MOEX), and news related with the company were chosen. To determine the dependencies between the news and the movements of shares quotes, a decision to use machine-learning technologies was made. During the execution of the work, the following tasks were set and achieved. There were collected the news from different sources. The obtained news contain the various data about the Yandex company, such as financial reports, press releases, information about different deals of the company and some other data. For news obtaining there was created a special «module», written in Python, which automates the process of news collecting. There was also created another module for news «post processing» to the form, which is needed for machine-learning algorithms. Along with the news, another special developed module collected the quotes information about Yandex shares. After that, all obtained data for the company was saved to PostgreSQL database on the remote server. Finally, four different machine-learning algorithms were trained and tested on the collected data and the best-performing model was chosen. As a result, the chosen machine-learning model has found the impact of news on stock market quotes of Yandex Company. However, the impact is quite insignificant. This means that the news have influence on stock market quotes, but the influence of the selected news is implicit. Nevertheless, the performed work shows the main principle, which can be applied for revealing the impact of news on stock market quotes using machine-learning models. In addition, this work will be useful in the future for more detailed study of the raised issue with using other methods of analysis like sentiment analysis of the news or with analysis, which use financial values as factor variables.