Optimization Of Parameter Support Vector Machine (SVM) using Genetic Algorithm to Review Go-Jek’s Services

Abstract

SVM method can be utilized to classify opinion data based on assessment attributes to discriminate whether an opinion is classified as positive or negative sentiment. The advantages of SVM implementation such as generalization, its stability, classification, and ability to process both linear and non-linear data have made SVM being considered as a reliable classification method. According to these considerations, the authors carried out a sentiment analysis using SVM algorithm to evaluate Go-Jek service reviews. This sentiment analysis is expected to provide benefits for the stakeholders, particularly for Go-Jek. Support Vector Machine (SVM) algorithm is capable of overcoming high dimensional data set, classification problems, and both kernel linear and non-linear regression, which make SVM reliable for an algorithm for classification and regression. However, support vector machine has a drawback in selecting relevant parameters to enhance its optimization. A genetic algorithm method is necessary to resolve the problem of selecting relevant parameters in support machine vector method.