Identifying Relative Contribution of Selected Technical Indicators in Stock Market Prediction

摘要

This paper investigates a sensitivity analysis of key technical indicators used in an artificial neural network (ANN) for predicting stock market trends. This process of forecasting future price movements in the stock market involves technical analysis. The data to be collected must have low noise or errors which would otherwise cause the production of inaccurate solutions/forecasts. This paper focuses on the data input selection based on the relative contribution of selected technical indicators while observing the strength of prediction and thereafter selecting a final set of inputs. Three different ANN architectures are compared along with relative weights of key technical indicators. Both the ability of each ANN model to predict stock market trends and a comparison of the contribution of indicators is determined. The models are then fit to real-world financial data over a five-year period. Sensitivity Analysis was used to extract the cause and effect relationship between the inputs and outputs of the network to improve the model's efficiency.