Switching Strategy of Recommendation Algorithms in Online Dating Platform
Renzhe Fang, Xingfa Shen, Yan Guo, Jian Yao, Jianhui Qiu · 2019
With the popularity of online dating platform, more and more platforms want to recommend suitable candidates to users. However, the recommendation algorithms they used do not meet accuracy, coverage, and other requirements at the same time. Therefore, using many recommendation algorithms with a certain weight to mix together which often increases the difficulty of project architecture and causes some unnecessary losses in the recommendation process. This paper proposes a set of strategies for switching recommendation algorithms in different situations with a certain rule, which is compatible with the advantages of various recommendation algorithms, reduces the overhead caused by mixed recommendation results in a single recommendation, and facilitates the personalized diversity of the algorithm. This paper compares the recommendation system under multiple switching strategies with the individual recommendation algorithm, analyze its advantages and disadvantages, and draw a conclusion that several switching algorithms are generally higher in accuracy and coverage.