A Comparative Analysis on the Machine Learning based Stock Price Prediction Models

Stock price estimation relies on analysing historical data together with market patterns to foresee forthcoming stock value. Equities demand stock price prediction to assist investors in their investment decisions. The traditional Stock market prediction methods that relied on ARIMA and Linear Regression performed poorly since the stock market exhibits unpredictable behaviour and complex patterns. The models currently face challenges when it comes to analysing non-linear patterns and noisy data because abrupt market changes reduce their reliability Recurrent neural networks (RNN) and convolutional neural networks (CNN). The study demonstrates that LSTM-based algorithms outperform other models in terms of predictive performance due to their capacity to manage and maintain historical market trends. Time-series patterns become detectable to these models because they process extensive datasets efficiently. A web application based on Flask technology enables users to access predictions in an easy-to-use manner. Our proposed system provides improved stock market prediction than traditional methods. The proposed stock marketing system becomes more effective in accuracy and usability than conventional methods.