We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
EŞİK DEĞERİNİN HİSSE SENEDİ FİYAT TAHMİN PERFORMANSINA ETKİSİNİN İNCELENMESİ.
- Authors
ÖZÇALICI, Mehmet
- Abstract
Artificial Neural Network models are successfully applied in the field of stock price predicting. Bias values are important factor that can affect the performance of neural networks. Although its importance, the effect of bias on stock price forecasting has not been investigated in literature. The purpose of this study is to examine the effect of bias on stock price predicting performance. For this purpose, historical price and volume information of stocks listed in BIST100 Index is used. By means of these data, 201 technical indicators are calculated. Activation function type in forecasting model and number of neurons in the hidden layer and variable selection are optimized with the Harmony Search Algorithm, a population-based meta-heuristic optimization method. RMSE and hit rate measurements are used as performance indicators. As a result, no statistically significant performance difference was found between biased and un-biased neural network models. However, it has been determined that the training of the models for un-biased models has been completed in a shorter time.
- Subjects
STOCK prices; ARTIFICIAL neural networks
- Publication
Hacettepe University Journal of Economics & Administrative Sciences / Hacettepe Üniversitesi Iktisadi ve Idari Bilimler Fakültesi Dergisi, 2017, Vol 35, Issue 4, p97
- ISSN
1301-8752
- Publication type
Article
- DOI
10.17065/huniibf.372406