Research On Key Technologies Of Internet Hot Topics Detection

With the Internet technology development and computer technological innovation, the Internet has evolved into Web2.0from the original form of the Web. Increasing number of new network resources and web applications live with "Island of infomation" and "Infomation overload" phenomenon. Discover and analysis hot topics from massive internet infomations become an important problem to be resoled. Although machine learning, natural language process technology has been widely used in network hot topic discovery, but the existing network hot topic detection algotithm still have relative limitations. The result of algorithm still cannot meet the customer’s standard, many issues there to be studied further. This study focus on the diversity and redundancy information resource, potential keywords, high-dimensional disater, time latency and other issues, provide a new method of resource aggregation, a new method of network information resource integration, a new method of keywords extraction based on Mantaras distance optimization and also an iteractive adaptive clustering algorithm based on ant colony optimization and hot topic filtering algorithms based on feature. High-performance and high-precision of the proposed algorithms were evalued by experiments.