1Department of Agricultural Economics, UAS, GKVK, Bengaluru- 560 065, Karnataka
2Dasceq Data Analyst Company, Bengaluru - 560 070, Karnataka
*Corresponding author: akshatakn@gmail.com
Online published on 22 February, 2021.
The classification of agricultural data is an important application of information technology in agriculture. Support vector machine (SVM) is a powerful state-of-the-art classifier and has been applied in many fields. In this paper, SVM is introduced to classify the agricultural data for improving the classification performance and forecast the data. In order to understand SVM, a study was conducted on credit cards of defaulter and non-defaulters using 12391 observations and 31 variables over 15 days. To measure the accuracy, the SVM model was compared with other classification techniques such as Decision Tree (DT), Random Forest (RF), and Logistic Regression (LR). The results highlighted that SVM (linear) has greater accuracy as comparison to other methods. SVM is an alternative, promising technique compared to time series forecasting. Another study on forecasting of oilseed production in India was done for the period from 1949 to 2017. The SVM model was compared with ARIMA model. The results highlighted that SVM model is better in comparison to ARIMA considering the mean absolute percentage error.
