According to the capability of analytical network process (ANP) in analysis of different dependencies and feedback relationships among elements of a decision problem, the current research aims to develop an ANP based method for the benchmark classification problems. Since the essential limitation of ANP is the increase of inconsistency in judgment of decision makers along with increase in problem dimensions, genetic algorithm is used to optimize ANP parameters and improve classification accuracy.
Considering the objective, this study is a developmental research and in term of data analysis, it’s a quantitative and mathematical modeling one. In this research, first a multi criteria decision making problem is developed based on ANP and in form of a classification problem and then the unknown parameters of a super matrix were calculated by machine learning methods. Next, the most proper values of these parameters which include thresholds of each class and the applied coefficients in the super matrix are estimated based on sample’s benchmarks or data.The following processes have been conducted througha genetic algorithm. Finally, in order to validate the proposed method, its performance is compared to some frequently used classification methods in the reviewed literature.
The results indicate the very competitive performance of the proposed method compared to known machine learning methods.
Multi-criteria Decision Making Methods (MCDM) are usually used for ranking purposes, however little attention has been paid to their high capabilities. In this paper ANP in combination with genetic algorithm demonstrated an efficient and suitable method in the field of data classification
- دسترسی به متن مقالات این پایگاه در قالب ارایه خدمات کتابخانه دیجیتال و با دریافت حق عضویت صورت میگیرد و مگیران بهایی برای هر مقاله تعیین نکرده و وجهی بابت آن دریافت نمیکند.
- حق عضویت دریافتی صرف حمایت از نشریات عضو و نگهداری، تکمیل و توسعه مگیران میشود.
- پرداخت حق اشتراک و دانلود مقالات اجازه بازنشر آن در سایر رسانههای چاپی و دیجیتال را به کاربر نمیدهد.