Comparison of machine learning algorithms using e-commerce data

Ever-evolving technology has been reflected in many areas, such as commerce, marketing, and businesses have had to capture innovation to avoid falling behind their competitors. This study compared machine learning algorithms using data from a website that continues to operate in Europe. The algorithms in question are;Naive Bayes is Support Vector Machines, K-MEANS, kNN, Regression. When the algorithms used in the study are examined with the large dataset, the highest success rate is Support Vector Machines, while Naive Bayes, kNN and Logistics Regression follow, respectively. In this study, the results were evaluated with accuracy, precision, sensitivity, kappa value and auc in this study using the Haldout method. Of the class balancing methods, smote, ros and rus were used. In the big data used in the study, it was observed that the algorithm that gave the slowest results was support vector machines.