Analyzing the behavior of internet customers based on social engineering

Background: The customers' opinions about the features and experience of using the products are considered as a valuable and reliable source for comparison and decision-making. Thus, the present study was an attempt to analyze the behavior of Internet customers based on social engineering. Methods: This study is applied research in the area of social networks. The statistical population of this study included Amazon social network users. The data includes XML and txt files brought to the programming environment. To analyze the behavior of Internet customers, a method based on the ensemble learning technique was implemented in MATLAB software. The common criteria that were used in data mining applications such as accuracy, sensitivity, and F-score. Results: The proposed model compared to other ensemble methods (support vector machines, Naive Bayes, ensemble neural networks, and decision tree ensemble) is in the priority in all three criteria for recognizing real and non-real users and has a better function. This method had high accuracy, precision, sensitivity, and F-criteria compared to other methods and it has a good status in evaluation criteria. The performance of the proposed model was much better than single algorithms and is the priority in terms of data mining evaluation criteria, but the training time for this model was much longer than other methods. Conclusion: The use of the proposed model in any organization that provides a product or service online, is quite promising and better results can be achieved with more studies.