Abstract

The aim of this research paper is to introduce the MLTrader trading strategy, which is one that utilizes machine learning techniques in order to make informed buy or sell decisions within a diverse stock portfolio. The cornerstone of the strategy are the sentiment analysis in financial news headlines that allow for detecting market sentiments and trends. Lumibot library, characterized by strong financial data analysis capabilities, was used for purposes of implementing it while it was also subjected to rigorous backtesting using Alpaca paper trading API. With libraries, modules and a custom MLTrader class integrated into it, one would consider MLTrader as being comprehensive. When these research findings are considered, they point to the effectiveness of this strategy in maneuvering through a dynamic stock market landscape with some promising results for its practical use in real world trading scenarios. In addition, this paper also contributes towards the emerging field of machine learning in finance by presenting an actionable and feasible strategy that could be further investigated and improved upon in relation to algorithmic trading methodologies.