Hengame Abbasi Mesrabadi - Department of Computer, Electrical Engineering, Zanjan Branch, Islamic Azad University, Zanjan, Iran
Ahmad Abdollahzadeh Barforoush - Artificial Intelligence Systems Laboratory, Department of Computer Engineering & Information Technology, Amirkabir University of Technology, Tehran, Iran,

Measuring recommender systems have become an important issue for building recommender systems. There is no specific methodology for measuring them yet. Since recommender systems are used in various domains, there are very challenging when measuring them. Thereby an effective and reliable methodology for measuring recommender systems is lacking. This paper presents a methodology for measuring recommender systems, conducting requirements engineering. The proposed methodology consists of 6 layers. These layers are: Requirements Classification, System Type, Quality Model, Metrics, Measurement and Indicator. In this work, requirements engineering contribution to methodology layers is introduced; a new structure for functional and nonfunctional requirements that is based on ISO-9126 model is created; all measuring metrics which can cover the recommender systems are defined. In conclusion, we evaluate our proposed methodology.

Requirements Engineering, Recommender System, Methodology, Measurement, Metrics, Quality