An AI-Enabled Stock Prediction Platform Combining News and Social Sensing with Financial Statements

Abstract

In recent years, the area of financial forecasting has attracted high interest due to the emergence of huge data volumes (big data) and the advent of more powerful modeling techniques such as deep learning. To generate the financial forecasts, systems are developed that combine methods from various scientific fields, such as information retrieval, natural language processing and deep learning. In this paper, we present ASPENDYS, a supportive platform for investors that combines various methods from the aforementioned scientific fields aiming to facilitate the management and the decision making of investment actions through personalized recommendations. To accomplish that, the system takes into account both financial data and textual data from news websites and the social networks Twitter and Stocktwits. The financial data are processed using methods of technical analysis and machine learning, while the textual data are analyzed regarding their reliability and then their sentiments towards an investment. As an outcome, investment signals are generated based on the financial data analysis and the sensing of the general sentiment towards a certain investment and are finally recommended to the investors.

Using Data Mining in the Sentiment Analysis Process on the Financial Market

“…In order to predict the future behavior of the stock prices and to consider forecasting a profitable behavior of the investor, in (Theodorou et al, 2021) additional signs of the stock behavior, in relation with the daily sentiment analysis were obtained.…”

Increasing Profitability and Confidence by using Interpretable Model for Investment Decisions

“…Moreover, there is a need to cover the semantic relationship between the aggregated news over time for more accurate and comprehensive results. (Theodorou et al, 2021) developed a platform called ASPENDYS to analyze daily stock data based on technical analysis as well as to consider news text reliability and social media sentiment concerning stocks. They provided a high-level interactive interface offering end users the functionalities of monitoring, modifying, and expanding their portfolio.…”

“…Although the authors focused on predicting mergers and acquisitions in financial companies with their proposed framework, however, they did not test it for canceled or pending deals. Additionally, relying solely on the positivity of actual news to determine the likelihood of a deal being completed may not adequately capture the complexities of real market dynamics (Theodorou et al, 2021).…”

Analytical Modeling and Empirical Analysis of Binary Options Strategies

“…Ref. [17] combines the news and social sensing with historical stock price to enhance the trading forecast.…”