• Riyanka hazra
• SUKDEB SAHA
• SINTHIA GOLDER
• MOULI MANDAL
• ZULFAQUARUL HAQUE

This study presents a comprehensive approach to predicting stock prices using Long Short-Term Memory (LSTM) neural networks, a type of recurrent neural network well-suited for time series forecasting. The dataset comprises historical stock prices from January 1, 2012, to the current date, obtained via the yfinance API. The primary goal is to develop a predictive model capable of forecasting future stock prices based on past data trends.Initially, the stock data was preprocessed to ensure suitability for model training. This involved calculating and visualizing key technical indicators such as the 100-day and 200-day moving averages of the closing prices. These moving averages are fundamental in financial analysis, providing insights into the stock's price trends and helping to identify potential buy or sell signals.Subsequent steps involved scaling and reshaping the data into sequences suitable for LSTM modeling. Data normalization was performed using the MinMaxScaler to scale the closing prices to a range between 0 and 1, ensuring that the model training process was efficient and effective. The data was then split into training and testing sets, with sequences of 100 days used to predict the subsequent day's stock price. The LSTM model was constructed with an architecture comprising an LSTM layer with 50 units, followed by a dropout layer to prevent overfitting, and a dense output layer with a single unit. The model was compiled using the Adam optimizer and the mean squared error loss function, both of which are standard choices for regression tasks in neural networks.Training the model involved fitting it to the training data for 50 epochs with a batch size of 32. This iterative process aimed to minimize the loss function and improve the model's predictive accuracy. The model's performance was evaluated on a separate test dataset, which was also preprocessed similarly to the training data.The evaluation metrics included the Root Mean Squared Error (RMSE), which quantifies the differences between predicted and actual values, providing a measure of the model's accuracy. Visual comparisons between the predicted and actual stock prices were also made to qualitatively assess the model's performance.The results demonstrated that the LSTM model could capture the underlying trends in the stock price data, making it a viable tool for stock price prediction. However, the study acknowledges the inherent complexities and unpredictability of financial markets, suggesting that while LSTM models can provide valuable insights, they should be used in conjunction with other analytical methods and domain expertise.This research underscores the potential of LSTM networks in financial[11] time series forecasting, contributing to the growing body of literature on machine learning applications in finance. Future work could explore the integration of additional features, such as trading volumes and other financial indicators[12], to further enhance the model's predictive capabilities. Additionally, the impact of varying model hyperparameters and the use of different neural network architectures could be investigated to optimize perform[1].

Stock Price Prediction·LSTM Networks·Moving Averages· Feature Engineering · Technical Indicators·Neural Networks·AI in Finance·Financial Forecasting

"Predictive Analytics for Stock Market Trends", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.11, Issue 7, page no.c238-c250, July-2024, Available :http://www.jetir.org/papers/JETIR2407231.pdf

"Predictive Analytics for Stock Market Trends", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.11, Issue 7, page no. ppc238-c250, July-2024, Available at : http://www.jetir.org/papers/JETIR2407231.pdf