Abstract:
Nowadays, stock analysis is an important task in financial market. Stock analysis methods including fundamental analysis and technical analysis are commonly used among financial professionals to help them on investment. In recent days, AI-based system becomes a common tool to predict stock price. Among AI-based models for stock forecasting, Artificial Neural Network (ANN) is the most popular and accurate model.
This research study has proposed data preprocessing by using relative movement to improve performance of ANN-based stock forecasting. Both fundamental and technical indicators are chosen as input to the system. The common preprocessing including Principal Component Analysis (PCA) and Z-Scaling are also applied. The evaluation metrics includes Root Mean Squared Error (RMSE), hit ratio, and total return. The k-fold cross validation is used to utilize the dataset of stocks in banking sector. The significance of those three metrics is determined through t-test over cross validation. The experiments show the promising result, in favor of the proposed model. The proposed model outperforms the traditional model, random walk model, and buy & hold strategy for all evaluation metrics.