CAJ | 학술논문

电子商务推荐虽然在促进电子商务发展过程中起到重要作用，但面临新商品推荐问题，新商品推荐问题的症结在于缺乏消费者与新商品之间的联系。从消费心理学视角，根据消费心理决定消费行为这一基本原理，以消费性格理论为指导，以贝叶斯网络为工具，以消费者的实际购物数据为基础，提出基于消费者购物记录分析其消费性格、基于消费者消费性格进行新商品推荐的方法，以消费心理为纽带建立消费者与新商品之间的联系。研究结果表明，该方法在统计学意义上具有可行性、合理性和一般性，即可以从消费者的购物记录中分析出其消费性格，能够据此判定某一新商品是否符合其消费性格。消费性格的引入把电子商务推荐中消费者偏好研究深入到消费心理层面，为发展出新的基于消费者消费心理的推荐方法提供借鉴。
전자상무추천수연재촉진전자상무발전과정중기도중요작용，단면림신상품추천문제，신상품추천문제적증결재우결핍소비자여신상품지간적련계。종소비심이학시각，근거소비심리결정소비행위저일기본원리，이소비성격이론위지도，이패협사망락위공구，이소비자적실제구물수거위기출，제출기우소비자구물기록분석기소비성격、기우소비자소비성격진행신상품추천적방법，이소비심리위뉴대건립소비자여신상품지간적련계。연구결과표명，해방법재통계학의의상구유가행성、합이성화일반성，즉가이종소비자적구물기록중분석출기소비성격，능구거차판정모일신상품시부부합기소비성격。소비성격적인입파전자상무추천중소비자편호연구심입도소비심리층면，위발전출신적기우소비자소비심리적추천방법제공차감。
Nowadays , recommendation systems in e-commerce are booming because of their potential and applicability for person-alized services to customers .However , recommendation is not always what customers are expecting , especially when new com-modities emerge in substantial quantities and buyer′s consuming styles become more diverse with the development of e-commerce. Yet there are more and more commodities not purchased or even never browsed , resulting in that less information could be used in recommendation .This was so-called the new commodities problem .The new commodities problem highlights the dilemma of e-recommendation .While e-recommendation necessarily needed because of the abundant commodities , yet on the other hand , new commodities lack the access to consumers , which traditional e-recommendations highly rely on .So it is difficult to recommend new commodities to consumers reasonably . The biggest problem of new commodities recommendation lies in there lacks the access to sufficient information linking consumer and new commodities .To cope with this problem , this paper proposes a new idea to establish the linkage within consumers and new commodities by introducing consumer psychology theories into e-recommendation .It had been manifested that consuming psychology is the causal relationship between a consumer and the commodities a consumer has selected or would select （ including new commodities）,i.e., the consumer′s psychology is a stable link between a consumer and so called commodities .If a consum-er′consuming psychologies were found out , then the linkage between a consumer and new commodities could be established .In this paper , the theory of consuming personalities is therefore combined with the Bayesian theory to build a new method of new commodities recommendation , which is mainly composed by two parts .One is the Bayesian network learning and the other is Bayesian network reasoning .Firstly, a consumer′s consuming personality is discovered from his /her consumption records by u-sing Bayesian network learning , then the trained Bayesian network is used to judge whether a new commodity is in conformity with the consuming personalities .The verification through the actual purchasing data shows that the built method has the properties of feasibility, reasonability, and generality.That is, consumers′consuming psychology could be revealed from one′s purchasing re-cords, and it can be used to infer whether a new commodity fits with one′s consuming personality . The method of new commodities recommendation is based on causal relationships between a consumer and new commodities , which is essentially different from the traditional e-recommendations based on correlation .The method not only deepens the user′s profile by introducing the consuming personalities into e-recommendation , but also provides the probability to develop a new recommendation method based on a consumer′s psychology .