Development of Machine Learning Algorithms for Predicting Price Movements in Financial Markets

EasyChair Preprint 15305

Abstract

In today’s volatile financial markets, accurately predicting stock prices is a monumental challenge.
Yet, with the surge in machine learning techniques, new doors have opened for tackling this complex
task. This paper dives deep into the performance of several machine learning models—Random
Forest, SVM, XGBoost, ARIMA, and LSTM—in predicting short-term stock price movements. Our
focus is on forecasting the adjusted close prices of major indices, including the S&P 500, NASDAQ,
and Dow Jones, using historical data spanning over two decades.
By meticulously engineering features like moving averages, volatility, and momentum indicators, we
aim to capture subtle market trends. We further enhance model performance through rigorous data
preprocessing and train-test splits to ensure robust evaluation. In our results, deep learning models
like LSTM outperform traditional models, demonstrating superior accuracy, especially in handling
market volatility. The findings underscore the potential of LSTM for real-time trading strategies,
positioning it as a powerful tool for short-term financial forecasting.

Keyphrases: AI, Finance, Investing, Predicting, Traiding, forecast, machine learning, neural network

@booklet{EasyChair:15305,
  author    = {Baiel Kurstanbek},
  title     = {Development of Machine Learning Algorithms for Predicting Price Movements in Financial Markets},
  howpublished = {EasyChair Preprint 15305},
  year      = {EasyChair, 2024}}