Contextual User Interest Modeling Approaches for Personalized Services on Multiple Devices

Nowadays, personalized services are quite pervasive and become more and more important. Many companies provide personalized services such as an item recommendation, a personal advertisement, etc. However, current personalized services have overlooked an important point that a user's focused interest is changing based on the contexts. For example, a person has an interest in outdoor sports when the weather is good and in online shopping if it's raining heavily. For sophisticated personalized service in these situations, this paper presents a novel and creative device-side approach to model user's contextual interests and to rank user's interests based on contextual differences. Based on this approach, a personalized service such as a personalized search could be performed. As a result, the smart device intelligence and user experience can then be greatly enhanced.