A Method for Estimating User’s Preference about Shopping Items Based on User’s Behaviors on Smartphone

抄録

Nowadays, with the wide spread of smart phone, many users turn to use their smart phones frequently on online shopping instead of immovable devices such as desktop PCs. For providing high quality service, the user’s evaluations on shopping items are useful for recommendation and search of items that the user would prefer to. However, it is not easy to obtain user’s manual evaluations on items. It is not convenient for user to make evaluations manually on each item. In this paper, we propose a simple and effective idea to estimate user’s preference on online shopping items that the user has already browsed without manual evaluation, based on the user’s browsing behaviors. This is performed under the hypothesis that the user’s preference of browsed items is reflected by the operation behavior on the smart phone. This hypothesis is verified by experiment, using our original application. The experimental results show the correlation between browsing behaviors and evaluation of product items by using statistical analysis. In addition, by using machine-learning techniques, the results show that the user’s preference of browsed item can be estimated effectively based on the user’s browsing behaviors.