A recommender system based on collaborative filtering, graph theory using HMM based similarity measures

Anshul Gupta () and Pravin Srinath ()
Additional contact information
Anshul Gupta: NMIMS University
Pravin Srinath: NMIMS University

International Journal of System Assurance Engineering and Management, 2022, vol. 13, issue 1, No 54, 533-545

Abstract: Abstract Collaborative filtering (CF) is a widely used method in recommendation systems (RS). In the CF, similar users' interests and preferences are analyzed and items that they might be interested in are recommended. However, the RS suggests items without maintaining the proper order (sequence) assuming that in the next purchasing user may select any single item from the list of recommended items. This limits the scope of the RS to a single future event, which does not match with the natural behavior of the users who generally purchase multiple items together in a specific sequence. Therefore in this paper, a more realistic RS is proposed which suggests a list of items in a proper sequence that users may purchase together. Since making such RS also required to measure the similarity among the users considering their sequential (temporal) purchasing behavior. The temporal similarity among the users could be calculated using a graph-based approach. However, this could lead to an NP-complete computational problem. Hence, to overcome this limitation, we also proposed an HMM-based approach that utilized the EMISSION Matrix to estimate the sequential similarity among the users. The EMISSION Matrix used with the proposed algorithm is a part of HMM (Hidden Markov Model). In HMM the EMISSION matrix is used to estimate the possibility of emission of a particular symbol from a process. In the presented work, the process is mapped with the user and the symbol with the selected item. Matching the EMISSION matrix of users' item selection process provides an improved similarity measure which also includes the sequential preferences. The performance of the proposed approach is measured for different configurations using the MovieLens dataset and several evaluation metrics. The proposed approach also overcomes the limitations of the previous algorithm which restricts the prediction domain to the user’s browsing history only.

Keywords: Recommender system; Collaborative filtering; Similarity measure; Behavior prediction; Hidden Markov model (HMM); Graph theory (search for similar items in EconPapers)
Date: 2022
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://link.springer.com/10.1007/s13198-021-01537-6 Abstract (text/html)
Access to the full text of the articles in this series is restricted.

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:spr:ijsaem:v:13:y:2022:i:1:d:10.1007_s13198-021-01537-6

Ordering information: This journal article can be ordered from
http://www.springer.com/engineering/journal/13198

DOI: 10.1007/s13198-021-01537-6

Access Statistics for this article

International Journal of System Assurance Engineering and Management is currently edited by P.K. Kapur, A.K. Verma and U. Kumar

More articles in International Journal of System Assurance Engineering and Management from Springer, The Society for Reliability, Engineering Quality and Operations Management (SREQOM),India, and Division of Operation and Maintenance, Lulea University of Technology, Sweden
Bibliographic data for series maintained by Sonal Shukla () and Springer Nature Abstracting and Indexing ().