Discovering Consumers' Purchase Intentions Based on Mobile Search Behaviors

Search activity is an essential part for gathering useful information and supporting decision making. With the exponential growth of mobile e-commerce, consumers often search for products and services that are closely relevant to the current context such as location and time. This paper studies the search behaviors of mobile consumers, which reflect their customized purchase intentions. In light of machine learning, a probabilistic generative model is proposed to discover underlying search patterns, i.e., when to search, where to search and in what category. Furthermore, the predicting power of the proposed model is validated on the dataset released by Alibaba, the biggest e-commerce platform in the world. Experimental results show the advantages of the proposed model over the classical content-based methods, and also illustrate the effectiveness of integrating contextual factors into modeling consumers search patterns.