Use of text mining in prediction of financial performance: A research in BİST manufacturing industry

Along with the financial reporting approach that has changed due to economic and technological development, textual data as well as numerical data have been included in financial reports and other disclosures of companies. This has led to the problem of data stacks containing text data that cannot be analyzed by conventional methods. Starting from this point, this study focuses on the analysis of text data in financial reports and examines whether financial reporting strategies of companies are shaped according to financial performance. The main objective of the study is to develop practical analysis techniques that will enable the use of text data as well as numerical data in the analysis of financial performance of companies. For this purpose, it is sought to answer the question whether there is a relationship between financial performance and the text data in the annual reports. In addition, the study investigated whether there is information about the future financial performance of the companies in the explanations in the annual reports. The research method is text mining. As the research model, support vector machines algorithm, which is one of the data mining classification algorithms, was used. The data set of the research is the annual reports and financial performance indicators published by the companies in the BIST manufacturing industry sector in the 8 yearly period between 2010-2017 years. The results of the research show that there is a relationship between financial performance and text data in annual reports. On the other hand, according to the results of the research, it can be said that the explanations containing the text data in the annual reports contain information about the future financial performance of the companies. In addition, the results of the research show that text mining is a useful and practical method in the analysis of text data presented in financial reports.