A CLASSIFICATION MODEL FOR E-COMMERCE BUSINESS-BASED ON BASIC DATA ATTRIBUTES OF BUSINESS ENTREPRENEURSHIP TO OPTIMIZE THE PREDICTION OF E-COMMERCE BUSINESS TYPE IN THAILAND

Abstract

    The preparation for market change in e-commerce businesses to enable them to compete effectively and move towards a better direction has led to this research. The research objectives were: 1) to present the best classification model of e-commerce business-based on basic data attributes operations; 2) to analyze attributes that affect the optimization suitable for predictive e-commerce businesses; 3) to compare the effectiveness of the classification model with the feature analysis model using t-test and ANOVA statistics. The sample was 400 corporations engaged in e-commerce businesses in Thailand, selected using a content-valid questionnaire with a reliability index of 0.91, and questionnaire reliability scores from 0.72 to 0.98. Data was collected via postal mail, and data analysis employed data mining as a forecasting. The findings revealed that: 1) The support vector machine model is the best for predictive e-commerce business type in Thailand with an accuracy of 80.97% 2) The best model for the optimization of the classification is the decision tree model with an accuracy of 80.04% and 3) The features that influenced the effectiveness of prediction included total business assets, business operation period, total current employees, average annual business income, and current job positions Moreover, 4) the decision tree model and the support vector machine model demonstrated no significant difference in effectiveness and can be used interchangeably. The research indicates that the models are capable of accurate and precise predictions, and users can correctly rank the importance of data attributes, which is essential for the success of e-commerce business types.