Semnan University International Journal of Nonlinear Analysis and Applications 2008-6822 16 1 2024 07 01 Analysis of user behavior using hierarchical classification and fuzzy logic 23 36 8525 10.22075/ijnaa.2023.32461.4828 EN Mohammad Hossein Hosseini Department of Computer Engineering, Safadasht Branch, Islamic Azad University, Safadasht, Tehran, Iran Azadeh Al-Sadat Khodaei Department of Computer Engineering, Safadasht Branch, Islamic Azad University, Safadasht, Tehran, Iran Journal Article 2023 10 27 In today's age, the Internet has become the most important source of information and its supply for many goods and services due to its ease of use, wide range and high speed. The unique characteristics of the Internet and its complete superiority over other markets have led many organizations to expand their services and products in the Internet markets. Therefore, the Internet environment has become a very competitive environment for organizations. In the process of buying and checking user and consumer behavior, goods or services will be exchanged, and these goods or services can be information, physical or virtual products, and even emotions. Many kinds of research have been conducted in the analysis of users' behavior, but these researches were not highly accurate, therefore, in this research, an attempt was made to provide a solution for recommending items based on users' past behaviors to improve the accuracy of past methods. This proposed method can eliminate the noise in the data by using fuzzy logic and hierarchical analysis which is used for the phase of selecting the effective features and also works on more effective data to increase the accuracy. The tree proposed in this research has the lowest possible height and this causes the computational overhead to be greatly reduced. The proposed method was compared with several competing methods and the results indicate the superiority of the proposed method over the compared methods. https://ijnaa.semnan.ac.ir/article_8525_bfd9d2b06473737126dadce8edd6711a.pdf