Prediction of Global Petrochemical Product Trade and Determining the Position of Iran: The Network Approach

Petrochemical industry has important role in creating value added for gas and oil resources, especially for Iran. Predicting import and export countries plus the trade product type, is a great aid for the industry stakeholders for optimal trade planning. Besides, computational methods of social networks, has been used in several applications and different areas. The aim of this paper is using network analysis methods for the first time for global petrochemical product’s trade. Data extracted and pr eprocessed from UN website of commercial trade, for popular petrochemical products from 2017 to 2019. Also, link prediction methods utilized for predicting next year trade relations, each year based on the previous one. Evaluation was done using two methods; computational approach, and validating the results with available data. The best performance method with more than 90% percent of AUC, was Preferential Attachment, and comparing the results with real data performed accordingly. Based on findings, the most promising countries for import was Spain, Slovenia, Australia, Norway and Argentina, and most feasible relation for Iran is exporting ACETONE to Spain. Finally, the methods to increase the performance of the predictions were proposed.