A new adaptive algorithm for v-support vector regression with feature selection using Harris hawks optimization algorithm

Support vector regression, especially, v-support vector regression (v-SVR) has been applied in several real problems. However, it is usually needed to tune manually the hyperparameter. In addition, v-SVR cannot perform feature selection. Nature-inspired algorithms were used as a feature selection and as an estimation for hyperparameter. In this paper, the Harris hawks optimization algorithm (HHOA) is proposed to optimize the hyperparameter of the v-SVR with embedding the feature selection simultaneously. Experimental results, obtained by running on two datasets, show that our proposed algorithm performs better than other methods, in terms of prediction, number of selected features, and running time. In addition, the HHOA's experimental results confirm the efficiency of the proposed algorithm in improving prediction performance and computational time compared to other nature-inspired algorithms, which show case HHOA's ability to search for the best hyperparameter values and to select the most informative features for prediction tasks. Therefore the HHOA may likely be ideal for defining the data relationship between input features and the target variable as opposed to other algorithms. In other real applications this is highly effective in making predictions.