Predicting fresh fruit and vegetables import of OECD countries' by data mining methods

Although, Turkey is one of the world's most important fresh fruit and vegetable producer countries, it can not adequately reflect this potential to export. In today's world, since the competition in the international trade is increasing day by day, it is very important to evaluate the available opportunities and resources in the most effective way while determining the target markets. In addition, in order to be successful in export process, target market study should be performed correctly. On the other hand, to prepare a future production plan and to make strategic decisions in this direction, it is necessary to determine the purchase demands of the importer countries for the next period. OECD countries have a significant potential in Turkey's fresh fruit and vegetable exports. Therefore, in this study, an analysis is carried out to predict OECD countries' fresh fruit and vegetable import by country and product base. For this purpose, the attributes that influence imports of OECD countries are taken as "Period", "Country code", "Product code", "Inflation rate", "Population", "GDP", "5 year later prediction of GDP", "Ease of doing business index", "Open market index", "Dolar exchange rate", "Production" and "Labor statistics in agriculture". The performance values of XGBoost, Random Forest and ANN in predicting the import variable were compared and the result of XGBoost Algorithm shows better performance compared to other methods.