Abstract:

Stock management is a branch of business management and is a critical subject of management for all companies. In order for to succeed in supply chain management, they also need to pay attention to stock management. In order to emphasize stock management, this study has studied the Turkish stone coal sector and tested the predictive models of the stock amount, which is an important cost element. Thanks to these predictive models, the factors that determine the amount of stone coal stocks have been identified. The monthly statistical tables regularly collected by the Turkish Stone Coal Authority (TTK) in the installation stage of the models of the stone coal industry with T.C. The energy and natural resources ministry’s economic indicator data has been used. The work period was January 2010-2015 and modelling was carried out with monthly data for the Poverty Office. The poverty agency is preferred because the "stock amount" capacity is excessive and the stock data is not missing. 6 independent variables have been used in the models as a dependent variable with the "stock amount" affecting the stock amount in advance. Relationships between these variables were studied by Hendry’s “General-Specific” modeling method. The classification-type models for the variables determined in the obtained specific model, the "artificial nerve networks (YSA)" and the decision trees, with multiple regression models, compared the forecast performance of all models. According to the results obtained in the study, the "Armutçuk" agency in the stone coal industry has been addressed and the artificial nerve network model has been selected for the highest performance stock amount estimate.

Keywords: