Alireza Ashouri.Roudposhti - Department of Business Management, Islamic Azad University, Science and Research Branch, Tehran, IranFaculty member of Tarjomane Oloum Institute of Higher Education
Ehsan Allah Moradi - Department of Business Management, Islamic Azad University, Science and Research Branch, Tehran, Iran

The importance of recognizing visitors to online sales websites has increased. Because most of these clients become customers who buy or will buy different products without referring to the company. This has changed the traditional relationship between buyer and seller, which will affect the production process, quality, supply process. In the traditional way, the buyer has a face-to-face relationship with the seller, based on empirical knowledge and individual ability in the field of spirits, tastes, preferences and in case of repetition of buying behavior on the principle of loyalty, product preference, buyer justification, customer buying habits. Evaluated and the supply equal to this amount of knowledge was established. In other words, in the traditional method, the seller, with the knowledge of the customer's behavior, offered products and selected the goods for sale, which were obtained experimentally. But in the online supply method, due to the lack of direct communication between buyer and seller, it is not possible for sellers to discover and produce customer behavioral knowledge. In cyberspace, on the other hand, customers share a lot of information about the traditional approach with others. Dissatisfaction or dissatisfaction in certain aspects was reported by customers to close people such as family, neighbors and friends, which is not defined in Internet practices (Xu et al., 2011) . Matrix shows the performance of the corresponding algorithms. It is also used in observer learning. Each column of the matrix shows an example of the predicted value. If each row contains a real (correct) instance. The computational intelligence of the clutter matrix is used to determine the value of evaluation criteria such as accuracy, readability, and f1. Accuracy refers to "how many of the selected samples are correct" and accuracy refers to the concept of "how many of the correct samples are selected" (Aggarwal, 2018).

Learning algorithms, customer satisfaction, digital marketing, clutter matrix system, online shopping sites