Decision Support System For Smartphone Recommendation: The Comparison Of Fuzzy AHP and Fuzzy ANP In Multi-Attribute Decision Making.

Previous researches outlined the advantages of the Analytical Hierarchy Process (AHP) and Analytic Network Process (ANP) methods in solving Multi-Attribute Decision Making (MADM) problems. The advancement of the above methods was continually developed as an effort to cover up various weaknesses. Mainly related to the consistency and linguistic variables in translating the expert opinions. Thus, it initialized the emergence of Fuzzy AHP (F-AHP) and Fuzzy ANP (F-ANP). Due to the restricted operation of these algorithms in smartphone selection, this research attempted to investigate the effectiveness of both methods in providing the analysis of criteria weight, the final recommendation weight, the product recommendation weight, and the execution time in DSS-SmartPhoneRec application development. A survey of one hundred respondents of University students identified the dominant criteria in selecting the smartphone, namely price, Random Access Memory (RAM), processor, internal memory, and camera. Hence, five alternative products were then chosen as the appropriate smartphones’ recommendations based on the respondent’s preferences. As an automatic tool, a DSS-SmartPhoneRec application was built to analyze and compare between F-AHP and F-ANP methods in resolving the smartphone selection cases. It revealed that the level of consistency of criteria weight, the final weight of recommendation, and the weight that the product-based F-ANP was 40% greater than F-AHP. In terms of execution time, F-AHP had a shorter time than F-ANP. Meanwhile, the comparison of products recommendation from DSS-SmartPhoneRec and a manual test showed that F-ANP was 16% more in line with the respondents’ predilection. In a nutshell, the DSS-SmartPhoneRec administered the devote smartphone recommendations based on the user’s expectation. The comparison analysis furnished a learning outcome for the users in determining the appropriate MADM method tailored to the type of cases