A Methodological Evaluation of Machine Learning Models for Stock Price Predictions

Stocks are possibly the most popular financial instrument invented for generating wealth and are the center point of investment portfolios. In the current financial environment, the ability to own and trade stocks is no longer confined to seasoned professionals but extends to individuals from all walks of life. Amidst this accessibility, the task of predicting stock prices remains a perennial challenge, with profound implications for investment decisions and portfolio management. Even with this accessibility, stock price prediction is still a difficult endeavor that has a big impact on portfolio management and investment choices. The development in trading technology has opened up the markets so that today nearly anyone can own stocks. This study was implemented to analyze how the various machine learning algorithms are performing in stock price prediction of a company by identifying different variations and patterns in the historical data. The ML Algorithms implemented in this study are Linear Regression, Moving Average, K-Nearest Neighbor (KNN) and Prophet. This study analyzes their performance in predicting the stock price for two different companies. Central to our investigation is the comparison of these machine learning models across diverse datasets, encompassing the stock prices of two distinct companies. The goal of this study is to identify the advantages and disadvantages of each algorithm, clarifying their relative benefits and drawbacks in the context of stock price prediction through rigorous review and empirical study. With the widespread adoption of AI in the field of finance it is necessary to choose the right algorithm for accurate and better decision making which motivated undertaking this study to present this comprehensive review.