A novel hybrid model using teaching–learning-based optimization and a support vector machine for commodity futures index forecasting

The analysis and prediction of financial time-series data are difficult, and are the most complicated tasks concerned with improving investment decisions. In this study, we forecasted a financial derivatives instrument (the commodity futures contract index) using techniques based on recently developed machine learning techniques. These methods have been shown to perform remarkably well in other applications. In particular, we developed a hybrid method that combines a support vector machine (SVM) with teaching–learning-based optimization (TLBO). The proposed SVM–TLBO model avoids user-specified control parameters, which are required when using other optimization methods. We assessed the viability and efficiency of this hybrid model by forecasting the daily closing prices of the COMDEX commodity futures index, traded in the Multi Commodity Exchange of India Limited. Our experimental results show that the proposed model is effective and performs better than the particle swarm optimization (PSO) + SVM hybrid and standard SVM models. For example, the proposed model improved the MAE by 65.87 % (1-day-ahead forecast), 55.83 % (3-days-ahead forecast), and 67.03 % (5-days-ahead forecast), when compared with standard SVM regression.