Stock Market Closing Cost Prediction using Neural Network via Moving Average Features

In recent times, huge random variations have been observed in stock prices on a daily basis. This makes it difficult for investors to predict investment options, rendering stock market prediction an open field of ongoing research. This paper aims to predict stock market closing prices using a neural network (NN)based approach. The proposed methodology first validates the basic NN-based stock market prediction (SMP) approach. Then, the range of moving average features is varied to evaluate prediction performance. Mean Absolute Error (MAE) and error histograms are used as performance metrics. The proposed method demonstrates the capability to achieve an accuracy of 9 9 . 9 8 7 % for Microsoft data, and for the FB database, it achieves 9 9 . 6 3 %. A regression model is employed to evaluate prediction performance.