Given the popularity of Web news services, we propose a topic mining framework that supports the identification of meaningful topics (themes) from news stream data. News articles are retrieved from Web news services and processed by data mining tools to produce useful higher-level knowledge, which is stored in a content description database. Instead of interacting with a Web news service directly, by exploiting the knowledge in the database, an information delivery agent can present an answer in response to a user request. A key challenging issue within news repository management is the high rate of documents update. That is, since several hundred news articles are published everyday by a single Web news service, it is essential to develop incremental data mining tools to cope with such dynamic environments. To this end, we present a sophisticated incremental hierarchical document clustering algorithm using a neighborhood search. The novelty of our proposed algorithm lies in exploiting locality information to reduce the amount of computation while producing high-quality clusters. Other components of topic mining (e.g., learning topic ontologies) can be performed based on the obtained document hierarchy. Experimental results show that our proposed incremental clustering produces high-quality clusters, and topic ontology provides an interpretation of the data at different levels of abstraction.
展开▼
