A Classification Technique for Online Shoppers' Purchasing Intention

Online shopping has become a recent outburst using electronic devices that allow consumers to shop online, as there are factors that influence shopping intention. The factors include consumers' attitudes, subject norms, behavior control, perception of usefulness, considered risks, and trust. This study aims to classify the intent of online consumers' shopping. Subsequently, based on these measures, this classification can provide fundamental and prioritized suggestions for performance improvements in online shopping. The classification is performed by using four data mining algorithms, which are random forest (RF), decision tree (DT), support vector machines (SVMs), and k-nearest neighbors (k-NN). Classification of online shoppers has been implemented based on the online shopping behavior dataset and critical performance metrics, including accuracy, precision, recall, and F1-measure. Essential factors have been considered in the online shopping behavior dataset that was used. The dataset has been preprocessed to check missing values and change the data to an appropriate numerical type. Then, the classification algorithms are applied to extract the required results. The results of detecting consumer shopping behavior showed that RF has outperformed other algorithms and achieved an average accuracy of 90.24%. The average classification accuracies of the DT, SVM, and k-NN are 89.59%, 89.38%, and 87.06%, respectively. The result of this work is hoped to assist in identifying online shoppers' behavior categories in practical aspects.