Fuzzy AHP and TOPSIS in cross domain collaboration recommendation with fuzzy visualization representation

Abstract

Cross domain collaboration recommendation method is proposed by combining fuzzy Analytic Hierarchy Process (AHP), fuzzy Technique for Order Preference by Similarity to Ideal Solution (TOPSIS), and fuzzy network graph for interactive visualization method. Existing cross-domain recommendation tackles the problem of sparsity, scalability, cold-start and serendipity issues found in single-domain, therefore the combination of fuzzy AHP and TOPSIS with visualization method may be able to give decision makers a quick start to initiate cross-domain collaborations. The proposed method is applied to the DBLP bibliographic citation dataset that consists of 10 domains in the field of computer science. Results show that the combination of fuzzy AHP and TOPSIS enables decision makers to find several authors from across domains that consist of 2.2 million publications in less than 3 minutes. The combination method will be represented in fuzzy visualization technique for fuzzy data. The establishment of the cross domain recommendation will set a stage for efficient preparation for researchers who are interested to venture into other domains to increase their research competency.