DIGITAL LIBRARY ARCHIVE
HOME > DIGITAL LIBRARY ARCHIVE
< Previous   List   Next >  
A Study on the Effect of Network Centralities on Recommendation Performance
Full-text Download
Dongwon Lee (College of Social Sciences, Hansung University)
Vol. 27, No. 1, Page: 23 ~ 46
Keywords
Recommender Systems, Personalized Recommendation, Social Network Analysis, Network Centrality, Classification Model
Abstract
Collaborative filtering, which is often used in personalization recommendations, is recognized as a very useful technique to find similar customers and recommend products to them based on their purchase history. However, the traditional collaborative filtering technique has raised the question of having difficulty calculating the similarity for new customers or products due to the method of calculating similarities based on direct connections and common features among customers. For this reason, a hybrid technique was designed to use content-based filtering techniques together. On the one hand, efforts have been made to solve these problems by applying the structural characteristics of social networks. This applies a method of indirectly calculating similarities through their similar customers placed between them. This means creating a customer’s network based on purchasing data and calculating the similarity between the two based on the features of the network that indirectly connects the two customers within this network. Such similarity can be used as a measure to predict whether the target customer accepts recommendations. The centrality metrics of networks can be utilized for the calculation of these similarities. Different centrality metrics have important implications in that they may have different effects on recommended performance. In this study, furthermore, the effect of these centrality metrics on the performance of recommendation may vary depending on recommender algorithms. In addition, recommendation techniques using network analysis can be expected to contribute to increasing recommendation performance even if they apply not only to new customers or products but also to entire customers or products. By considering a customer’s purchase of an item as a link generated between the customer and the item on the network, the prediction of user acceptance of recommendation is solved as a prediction of whether a new link will be created between them. As the classification models fit the purpose of solving the binary problem of whether the link is engaged or not, decision tree, k-nearest neighbors (KNN), logistic regression, artificial neural network, and support vector machine (SVM) are selected in the research. The data for performance evaluation used order data collected from an online shopping mall over four years and two months. Among them, the previous three years and eight months constitute social networks composed of and the experiment was conducted by organizing the data collected into the social network. The next four months’ records were used to train and evaluate recommender models. Experiments with the centrality metrics applied to each model show that the recommendation acceptance rates of the centrality metrics are different for each algorithm at a meaningful level. In this work, we analyzed only four commonly used centrality metrics: degree centrality, betweenness centrality, closeness centrality, and eigenvector centrality. Eigenvector centrality records the lowest performance in all models except support vector machines. Closeness centrality and betweenness centrality show similar performance across all models. Degree centrality ranking moderate across overall models while betweenness centrality always ranking higher than degree centrality. Finally, closeness centrality is characterized by distinct differences in performance according to the model. It ranks first in logistic regression, artificial neural network, and decision tree with numerically high performance. However, it only records very low rankings in support vector machine and K-neighborhood with low-performance levels. As the experiment results reveal, in a classification model, network centrality metrics over a subnetwork that connects the two nodes can effectively predict the connectivity between two nodes in a social network. Furthermore, each metric has a different performance depending on the classification model type. This result implies that choosing appropriate metrics for each algorithm can lead to achieving higher recommendation performance. In general, betweenness centrality can guarantee a high level of performance in any model. It would be possible to consider the introduction of proximity centrality to obtain higher performance for certain models.
Show/Hide Detailed Information in Korean
네트워크 중심성 척도가 추천 성능에 미치는 영향에 대한 연구
이동원 (한성대학교)
Keywords
추천 시스템, 개인화 추천, 사회 연결망 분석, 중심성 척도, 분류 모형
Abstract
개인화 추천에서 많이 사용되는 협업 필터링은 고객들의 구매이력을 기반으로 유사고객을 찾아 상품을 추천할 수 있는 매우 유용한 기법으로 인식되고 있다. 그러나, 전통적인 협업 필터링 기법은 사용자 간에 직접적인 연결과 공통적인 특징 을 기반으로 유사도를 계산하는 방식으로 인해 신규 고객 혹은 상품에 대해 유사도를 계산하기 힘들다는 문제가 제기되어 왔다. 이를 극복하기 위하여, 다른 기법을 함께 사용하는 하이브리드 기법이 고안되기도 하였다. 이런 노력의 하나로서, 사회연결망의 구조적 특성을 적용하여 이런 문제를 해결하려는 시도가 있었다. 이는, 직접적으로 유사성을 찾기 힘든 사용자 간에도 둘 사이에 놓인 유사한 사용자 또는 사용자들을 통해 유추해내는 방식으로 상호 간의 유사성을 계산하는 방식을 적용한 것이다. 즉, 구매 데이터를 기반으로 사용자의 네트워크를 생성하고 이 네트워크 내에서 두 사용자를 간접적으로 이어주는 네트워크의 특성을 기반으로 둘 사이의 유사도를 계산하는 것이다. 이렇게 얻은 유사도는 추천대상 고객이 상품의 추천에 대한 수락여부를 결정하는 척도로 활용될 수 있다. 서로 다른 중심성 척도는 추천성과에 미치는 영향이 서로 다를 수 있다는 점에서 중요한 의미를 갖는다 할 수 있다. 이런 유사도의 계산을 위해서 네트워크의 중심성을 활용할 수 있다. 본 연구에서는 여기서 더 나아가 이런 중심성이 추천성과에 미치는 영향이 추천 알고리즘에 따라서 도다를수 있다는 데에서 주목하여 수행되었다. 또한, 이런 네트워크 분석을 활용한 추천기법은 신규 고객 혹은 상품 뿐만 아니라 전체 고객 혹은 상품으로 그 대상을 넓히더라도 추천성능을 높이는데 기여할 것을 기대할 수 있을 것이다. 이런관점에서 본 연구는 네트워크 모형에서 연결선이 생성되는 것을 이진 분류의 문제로 보고, 추천 모형에 적용할 분류 기법으로 의사 결정나무, K-최근접이웃법, 로지스틱 회귀분석, 인공신경망, 서포트 벡터 머신을 선택하고, 온라인 쇼핑몰에서 4년2개월 간 수집된 구매 데이터로 실험을 진행하였다. 사회연결망에서 측정된 중심성 척도를 각 분류 기법에 적용하여 생성한 모형을 비교 실험한 결과, 각 모형별로 중심성 척도의 추천 성공률이 서로 다르게 나타남을 확인할 수 있었다.
Cite this article
JIIS Style
Lee, D., , "A Study on the Effect of Network Centralities on Recommendation Performance", Journal of Intelligence and Information Systems, Vol. 27, No. 1 (2021), 23~46.

IEEE Style
Dongwon Lee, "A Study on the Effect of Network Centralities on Recommendation Performance", Journal of Intelligence and Information Systems, vol. 27, no. 1, pp. 23~46, 2021.

ACM Style
Lee, D.,, 2021. A Study on the Effect of Network Centralities on Recommendation Performance. Journal of Intelligence and Information Systems. 27, 1, 23--46.
Export Formats : BiBTeX, EndNote
Advanced Search
Date Range

to
Search
@article{Lee:JIIS:2021:832,
author = {Lee, Dongwon},
title = {A Study on the Effect of Network Centralities on Recommendation Performance},
journal = {Journal of Intelligence and Information Systems},
issue_date = {March 2021},
volume = {27},
number = {1},
month = Mar,
year = {2021},
issn = {2288-4866},
pages = {23--46},
url = {},
doi = {},
publisher = {Korea Intelligent Information System Society},
address = {Seoul, Republic of Korea},
keywords = { Recommender Systems, Personalized Recommendation, Social Network Analysis, Network Centrality and Classification Model },
}
%0 Journal Article
%1 832
%A Dongwon Lee
%T A Study on the Effect of Network Centralities on Recommendation Performance
%J Journal of Intelligence and Information Systems
%@ 2288-4866
%V 27
%N 1
%P 23-46
%D 2021
%R
%I Korea Intelligent Information System Society