A Fine Tuned Model of Grasshopper Optimization Algorithm with Classifiers for Optimal Text Classification

Text classification is widely used application of natural language processing and machine learning classifier. Tuning hyper-parameters of the classifier in machine learning is difficult and important step. The tuning of hyper-parameters means selection of the optimal hyper-parameters of the algorithm. Many approaches for tuning of the parameters have been tested and applied in the literature for the improvisation of the classification performance. Further, the feature selection using machine learning is a major task, dealing with high-dimensional dataset. In this paper, we proposed a hybrid model of tuned Grasshopper optimization algorithm with classifiers. Grasshopper optimization algorithm works on the mimic behavior of grasshoppers. The aim of this meta-heuristic approach is to determine the minimal feature subset from all features to improve the classification performance. For tuning of the classifiers, we have used random search technique. The classifiers chosen for classification are K-Nearest Neighbor and Support Vector Machine. Five multi-class datasets are used to evaluate the performance of the model in terms of Accuracy and AUC curve. All results are computed with 10-fold-cross validation method. The evaluated results of the proposed model is compared with other algorithms, which verifies the performance of our technique. The proposed model outperformed among all the compared state-of-the-art techniques.