A STOCK PRICE PREDICTION MODEL USING MACHINE LEARNING TECHNIQUES

Abstract

It is a complex business to forecast stock prices due to volatility in the financial markets. Unique predictions could be quite advantageous to buyers, traders and financial bodies. This model uses ML methodologies for the prediction of stock values through the review of ancient data and technical intimations. The data has stock prices, trading volumes, and indicators, SMAs, EMAs, RSI, and Bollinger Bands. Many algorithms are used for identifying styles and traits, such as Linear Regression, support Vector Regression (SVR), Random forest, and long short-term memory (LSTM) networks that are best characterized by their effectiveness in handling time-series data. The data is preprocessed with cleaning, scaling using MinMaxScaler, and feature engineering including lag features and quantity trends. During evaluation and training both the models are tested with parameters like MAE or mean absolute error, RMSE or root mean squared error and R² score to standardize predictive accuracy. Consequences demonstrate that the accuracy of LSTM and elastic internet models is at the highest, however other models also have strong performance, and they include SVR and random forest as well. The technology makes tremendous insights through visualizations analyzing expected as well as actual stock prices, hence improving decision making of financial trading methods. Future improvements may include spherical analysis and reinforcement learning optimization