The Role of Technical Indicators in the Intraday Prediction of Stock Markets: Artificial Neural Network Models for Borsa Istanbul

Document Type : Article


1 Istanbul Commerce University, Information Technologies Application and Research Center, Istanbul, 34445 Turkiye

2 Istanbul Commerce University, Faculty of Business, Finance and Banking Department, Istanbul, 34445 Turkiye


In this study, two simulation models have been developed to predict the main stock price index of Borsa Istanbul (BIST100) with an artificial intelligence approach. In order to analyze the role of technical indicators in intraday predicting of stock markets, two different artificial neural network models have been developed in which different parameters are defined in the input layers. In the first model, 5 input parameters have been defined as open price (OP), highest price (HP), lowest price (LP), and two different moving averages (MA), 3 more parameters added as The Relative Strength Index (RSI), The Moving Average Convergence Divergence (MACD) and the moving average of MACD (TRIGGER). 70% of the data used in multi-layer network models developed with a total of 97 data sets have been used for training the model, 20% for validation and 10% for testing. The results show that both ANN models can predict BIST100 values with very low error rates. However, it is seen that the prediction performance of the first model, which has been developed by defining fewer input data, is higher than the second model. The results obtained support that the predictions made with intraday data are stronger between 13:00 and 16:30.