Stock Recommendations Using Machine Learning and Natural Language Processing

Due to the increasing number of investors in the past decade, the financial markets are now accountable for any country’s economic stability. Stock markets have grown increasingly unpredictable in recent years, yet they continue to be the most crucial for both investors and industries. There are several stock trading advice systems on the market that claim to be able to accurately predict future trends. Recommendation systems for stock trading are of great significance to a layperson who wants to benefit from stock trading despite not having a seasoned trader’s capacity or experience. The present study proposed a three-fold function. First, it provides a comparative analysis of the existing recommendation systems. It then applies and compares the results of the four Machine Learning models for price prediction on a real dataset. Finally, six sentiment analysis models are compared for the analysis of stock-based tweets. The best of the two are lastly integrated to arrive at a stock buy or sell recommendation.