CAJ | 학술논문

网上营销的难点是如何寻找潜在客户。识别用户行为是一个复杂的过程，建议采用二维模型来寻找相关的项目，然后推荐给用户。但太多不相干的建议转化为用户信息过载问题。本研究的目的是在每个维度中找出已被用户搜索的项目／对象，最受好评的每个维度中的项目可给予建议。要做到这一点，我们采用张量，它可以有效地模拟多维Web日志中的用户搜索数据，然后确定在每一个搜索维度上的用户模型。所有的评估均基于真实数据集，这清晰地显示出比其他广泛使用的向量和矩阵方法的有效性。
망상영소적난점시여하심조잠재객호。식별용호행위시일개복잡적과정，건의채용이유모형래심조상관적항목，연후추천급용호。단태다불상간적건의전화위용호신식과재문제。본연구적목적시재매개유도중조출이피용호수색적항목／대상，최수호평적매개유도중적항목가급여건의。요주도저일점，아문채용장량，타가이유효지모의다유Web일지중적용호수색수거，연후학정재매일개수색유도상적용호모형。소유적평고균기우진실수거집，저청석지현시출비기타엄범사용적향량화구진방법적유효성。
How to find potential customers is the difficulty of online marketing. The paper recommends the using of two-dimensional model to find related items, and then recommend to the user. But too many irrelevant suggestions are translated into the overload problem of user information. The purpose of this study is to identify in each dimension the project / object that searched by users, and the most award-winning project in each dimension can give advice to user. The paper uses tensor to do this work, which can effectively simulate the multi-dimensional Web logs in user＇ s search data, then determine a search of user model in each dimension. All assessments are based on real data sets, which dearly show the effect than the widely used vector and matrix method.