Abstract: 【Purpose/significance】To solve the problems of low recommendation accuracy of data sparse users and low overall satis⁃
faction of big data alliance group users with the group recommendation results, this paper proposes a recommendation method of big
data alliance data resource group based on improved VIKOR.【Method/process】According to the characteristics of data resources
group users of big data alliance, users are divided into in-group users and out-group users when constructing group recommendation
matrix, and the influence of different user ratings on group recommendation results is considered respectively. According to the partic⁃
ularity of data resources of big data alliance, a method to determine the attribute weight of data resources is proposed to confirm the at⁃
tributes of different data resources, so as to improve the quality of group recommendation.【Result/conclusion】Experimental results
show that the algorithm proposed in this paper can not only provide accurate recommendation results for data sparse users, but also ef⁃
fectively improve the overall satisfaction of data resource group users of big data alliance.【Innovation/limitation】In this paper, the im⁃
proved VIKOR algorithm is used in the group recommendation of data resource of the big data alliance, which effectively improves the
effect of group recommendation, but does not consider the impact of user clustering on the results of group recommendation. Next, we
will study how alliance users are clustered.