Abstract
Detecting and understanding web search intention have considerably gained in interest during the last years. This is challenging in mobile environments due to the impact of ubiquity on the mining process: changing user’s context influences the information needs, resource constraints affect the process execution. Moreover, in mobile devices it is possible to find situations where the analysis of user’s local behavior is required. Then, it is necessary to calculate locally the model using only local data. The need of autonomy required in a situation in which the data miner is not present makes the problem even more challenging. In this paper we address the problem of query categorization in mobile devices. Hence, we present an approach for categorizing local user queries and analyze the main challenges associated to the problem of performing the data mining process on a mobile device. Validation of the approach is shown by applying incremental clustering over a real data set.