Özet:

Continuous or sudden changes in competitive conditions pose difficulties for businesses, stock investors and other actors in the transportation sector to make strategic moves. In this case, if the indices reflecting the upward and downward movements in the stocks can be interpreted and estimated correctly it will provide a more sustainable competitive advantage. Under that circumstances, the aim of the study was tried to be predicted with the BIST Transportation index (XULAS) Artificial Neural Networks (ANN) Nonlinear External Input Autoregressive Network (NARX) model, which includes the stocks of nine companies operating in the transportation sector. Accordingly, ten macroeconomic variables on a monthly basis and six on a daily basis, have been identified which have a direct or indirect effect on XULAS. In this study, two different models were suggested as monthly and daily. While the daily data used in the study covers the period of 19/07/2010-15/10/2021, the monthly data covers the period of 08/2010-08/2021. As performance criteria, the results of MAD, MSE, RMSE and MAPE values were examined. As a result of the prediction, it can be said that the proposed ANN NARX network models are quite successful in predicting the XULAS values. When both models are compared, it has been determined that working with daily data gives more successful results in terms of forecasting and prediction.

Anahtar Kelimeler: