Multi-attribute Based Inﬂuence Maximization in Social Networks Qiufen Ni 1 , Jianxiong Guo 2,3(B ) , and Hongmin W. Du 4 1 School of Computers, Guangdong University of Technology, Guangzhou 510006, China niqiufen@gdut.edu.cn 2 BNU-UIC Institute of Artiﬁcial Intelligence and Future Networks, Beijing Normal University at Zhuhai, Zhuhai 519087, Guangdong, China jianxiongguo@bnu.edu.cn 3 Guangdong Key Lab of AI and Multi-Modal Data Processing, BNU-HKBU United International College, Zhuhai 519087, Guangdong, China 4 Accounting and Information Systems Department, Rutgers University, Piscataway, NJ 08854, USA hd255@scarletmail.rutgers.edu Abstract. Viral marketing on social networks is an important applica- tion and hot research problem. Most of the related work on viral market- ing focuses on the spread of single information, while a product may asso- ciate with multi-attribute in real life. Information on multiple attributes of a product propagates in the social networks simultaneously and inde- pendently. The attribute information that a user receives will determine whether he would purchase the product or not. We extend the tradi- tional single information inﬂuence maximization problem to the Multi- attribute based Inﬂuence Maximization Problem (MIMP). We present the Multi-dimensional IC model (MIC model) for the proposed problem. The objective function for MIMP is proved to be non-submodular, then we solve the problem with the Sandwich Algorithm, which can get a max  f (S U ) f (S U ) , f (S ∗ L ) f (S ∗ o )  (1 - 1/e) approximation ratio to the optimal solu- tion. Experiments are conducted in two real world datasets to verify the correctness and eﬀectiveness of the proposed algorithm. Keywords: Social network · Inﬂuence maximization · Multi-attribute information · Approximation algorithm 1 Introduction Online social networks are an important class of graph data. Data by gener- ated by users have been growing rapidly through various online social net- works, such as Facebook, LinkedIn, ResearchGate, and messengers like Skype and WeChat [1]. The widespread use of these social platforms leads to an increas- ing interest in mining important and useful but implicit patterns. Eﬃcient tech- niques for extracting information from graph data are crucial to applications c  Springer Nature Switzerland AG 2021 W. Wu and H. Du (Eds.): AAIM 2021, LNCS 13153, pp. 240–251, 2021. https://doi.org/10.1007/978-3-030-93176-6_21