Along with the daily operation of e-commerce web services, a significant quantity of data has been recorded. The research of user's behaviors based on the collected data has generated intense attention for accurately offering services that can match the customer's needs and predict the purchase actions. Traditionally, most of the researches utilize only the behavioral instances between users and products, i.e., browse or click history, and session status. However, these features provide only a

more »

... undamental knowledge of the given user rather than the rationale behind their actions. We find that query should play an important role as well as it is the main entry point for users when arriving e-commerce website. Since users utilize queries to decide the direction of succeeding event, the semantic meanings of these queries demonstrate a particular link with the action. In this paper, we propose the Prediction framework that analyzes User's Sequential Actions via Context (PSAC) to exploit the connection between the user's searching keywords and behaviors to investigate their ultimate intention on an e-commerce website and improve the purchase prediction accuracy. We utilize the ecommerce dataset provided by Yahoo Taiwan, one of the largest web services provider in Taiwan. According to our preliminary analysis, we design a session-based structure to deal with the environmentshifting (influenced by coexisting fashion), and experience-shifting (changed through user's actions) issues which we observed in the dataset. In the simulation section, we apply two deep learning frameworks to perform the prediction task. Experimental results confirm that queries serve as a critical matter in perceiving a user's purchasing intention. Moreover, the proposed framework could significantly improve the prediction accuracy compared with baseline methods.