Indian Journal of Science and Technology
DOI: 10.17485/ijst/2019/v12i4/140756
Year: 2019, Volume: 12, Issue: 4, Pages: 1-21
Original Article
Adel A. Nasser1* , A. A. Alkhulaidi2 , Mansoor N. Ali3 , M. Hankal4 and M. Al-olofe4
1Department of Information Systems, Sa’adah University, Sa’adah, Yemen; [email protected]
2Department of Computer Science, Sana’a University, Sana’a, Yemen; [email protected]
3Department of Information Systems, Sana’a University, Sana’a, Yemen; [email protected]
4 Electricity Department, Sana’a University, Sana’a, Yemen; [email protected], [email protected]
*Author for correspondence
Adel A. Nasser
Department of Information Systems, Sa’adah University, Sa’adah, Yemen.
Email: [email protected]
Objectives: To study the impact of the seven normalization methods on the results of ranking using three ranking methods and based on four objective methods. It is,furthermore, analyzes the characteristics of eighty four integrated approaches for multi-attribute decision making (MADM) problems. Methods/Statistical Analysis: Definitely, the proposed combinations of ranking approaches are developed based on the TOPSIS, the Simple Additive Weighting and Weighted Euclidian distance ranking methods and based on the equal weight, Entropy, The Criteria Importance Through Inter-criteria Correlation and Statistical variance procedure objective weighting methods, through the implementation of seven normalization techniques’ namely the Weitendorf’s, Korth’s, Nikamp’s, Unitization, Quotient Transformation (Q-T) - R, Q-T mean, and Q-T-PSK normalization methods. Absolutely, the whole proposed approaches are applied to the field of healthcare decision making in Yemen respectively. The Pearson Correlation Coefficient (PCC), Change and the overall level of change in the positions of the alternatives are used to compare the impact of the selected normalization methods on the result of ranking obtained by all ranking approaches. Findings: The result shows that the normalization formula influences the result of the ranking using both TOPSIS and SAW methods, but the level of the impact is changed by the changing of the objective method. It also, offers that the choice of normalization formula doesn’t influence the result of the ranking using WED method. In this study, high and very high numerical correlation coefficients between places in the different rankings are observed. Deferent changes in the positions, and in the overall level of change in case of particular governorates are also observed. The normalization methods are ordered by its impact on the result of the ranking for each weighting method used with each ranking one. Application: The ordering of the normalization methods by its impact on the ranking’s results obtained using SAW and TOPSIS and based on deferent types of weighting methods must be taken into consideration while solving related MADM problem. And, the ordering results of these methods in our study give some recommendations for amending this problem.
Keywords: Euclidian Distance, Impact of the Data Normalization Methods, Weighted Euclidean Distance, SAW, TOPSIS
Subscribe now for latest articles and news.