An Efficient Model for Predicting Future Price of Agricultural Commodities using K-Nearest Neighbors Algorithm Compared with Support Vector Machine Algorithm

Abstract

The main idea of the proposed research work is an effective production plan for future price prediction of agricultural commodities using the K-Nearest Neighbors algorithm with novel hamming code over the Support Vector Machine learning algorithm. Materials and Methods: For predicting the future price of agricultural products, this research study looks at two algorithms: the K-Nearest Neighbors algorithm with novel hamming code and the Support Vector Machine technique. The sample size for each algorithm is 20, and G power is 80%. Results: On the dataset utilized, the K-Nearest Neighbors classifiers have a prediction accuracy of 60.67% whereas the Support Vector Machine technique has a prediction accuracy of 40.56% An independent sample T-test yielded the statistical significance P = 0.041 (P<0.05). Conclusion: The K-Nearest Neighbors algorithm obtained better accuracy when compared to the Support Vector Machine technique.