The effectiveness of the SVM method in the task of determining profitable organizations

The article consider various options of applying the method of support vectors (SVM) for solving problems in the field of agricultural production. In particular, the effectiveness of SVM is analyzed in determining factors affecting the profitability/unprofitability of organizations in the Nizhny Novgorod region engaged in agricultural production. The considered factors included the size of the organization and the amount of available economic resources. Using a recursive extraction procedure, five most decisive factors that affected the profit obtained in 2007–2017 were selected for each of the 12 types of agricultural products considered. For each subsample, models were built using SVM (with different cores and parameters), the most optimal among which was selected and its accuracy was evaluated. As a result, the considered factors were ranked according to their impact on the profitability/unprofitability of an organization, depending on the types of products produced.

Authors: A. D. Cheremuhin, A. A. Shamin, M. O. Kolbanev, V. V. Tsehanovsky

Direction: Informatics, Computer Technologies And Control

Keywords: support vector machine, agricultural organizations, feature selection, profitability, recursive extraction, rating, parameter optimization, model accuracy

---

View full article