TITLE:
Incorporating User’s Preferences into Scholarly Publications Recommendation
AUTHORS:
Tobore Igbe, Bolanle Ojokoh
KEYWORDS:
Personalization, Digital Library, Information Retrieval, Recommender System, Citation Analysis, User Preferences
JOURNAL NAME:
Intelligent Information Management,
Vol.8 No.2,
March
29,
2016
ABSTRACT: Over the years, there has been increasing
growth in academic digital libraries. It has therefore become overwhelming for
researchers to determine important research materials. In most existing
research works that consider scholarly paper recommendation, the researcher’s
preference is left out. In this paper, therefore, Frequent Pattern (FP) Growth
Algorithm is employed on potential papers generated from the researcher’s
preferences to create a list of ranked papers based on citation features. The
purpose is to provide a recommender system that is user oriented. A walk
through algorithm is implemented to generate all possible frequent patterns
from the FP-tree after which an output of ordered recommended papers combining
subjective and objective factors of the researchers is produced. Experimental
results with a scholarly paper recommendation dataset show that the proposed
method is very promising, as it outperforms recommendation baselines as
measured with nDCG and MRR.