Forecasting of market sentiments using artificial intelligence

Abstract

Every investor faces the challenge of making efficient investment decisions. There are many methods to analyze the causes of changes in the financial market and to predict future trends based on such information. One way is to predict investors sentiment. This type of study is not extensively studied in the scientific literature, so the purpose of this article research is to perform prediction of different investor sentiment and to evaluate the reliability of the model used for prediction, i.e. reaching to discover a reliable sentiment prediction algorithm. Artificial intelligence deep learning short-term memory (LSTM) network method and graphical representation of the obtained results are used for the research. The study found that the margin of error (RMSE) obtained for each sentiment prediction case was very low, which means that the algorithm used for prediction is very reliable. Investors using this algorithm can help themselves better study market trends, but other financial market research methods should be used in parallel to make investment decisions as efficiently as possible.