Scholarlink

This research presents an intelligent system for forecasting the stock market for the very next day by using a Hybrid approach of Machine Learning and Artificial Intelligence to process financial time series. The Overall system is based on two modules. The first module is identifying potential stocks by scanning all 50 stocks of Nifty 50, by checking for the highest volatility and positive correlation with Nifty Indices of NSE using a combined approach including ARCH and GARCH, which makes it more robust in terms of portfolio selection of best profit-making stocks. The second module will process all the selected resulting stocks in the EMA crossover momentum Strategy to check their upward or downward trend and predict their next day Open and Close prices to make the strategy for our portfolio to earn maximum profit. Predicting stock prices is based on an enhanced Auto-ARIMA model with the AIC method, which makes it different and more accurate from other traditional ARIMA approaches. Accuracy has also been highlighted by the error evaluation metrics of 3 models of ARIMA. Comparative analysis with traditional ARIMA models using standard error evaluation metrics confirms the improved accuracy and practical utility of the proposed system.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI-Driven Forecasting of Financial Time Series: A Hybrid Machine Learning Approach to Stock Market Prediction with Volatility Modelling

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Garima S. Makhija,
  • Mohammad Atique

摘要

This research presents an intelligent system for forecasting the stock market for the very next day by using a Hybrid approach of Machine Learning and Artificial Intelligence to process financial time series. The Overall system is based on two modules. The first module is identifying potential stocks by scanning all 50 stocks of Nifty 50, by checking for the highest volatility and positive correlation with Nifty Indices of NSE using a combined approach including ARCH and GARCH, which makes it more robust in terms of portfolio selection of best profit-making stocks. The second module will process all the selected resulting stocks in the EMA crossover momentum Strategy to check their upward or downward trend and predict their next day Open and Close prices to make the strategy for our portfolio to earn maximum profit. Predicting stock prices is based on an enhanced Auto-ARIMA model with the AIC method, which makes it different and more accurate from other traditional ARIMA approaches. Accuracy has also been highlighted by the error evaluation metrics of 3 models of ARIMA. Comparative analysis with traditional ARIMA models using standard error evaluation metrics confirms the improved accuracy and practical utility of the proposed system.