User Recommendation Based on Semantic Pattern

With the development of Search Engine, more and more people obtain information and materials easily. However, it is not convenient and quick to get information because we have to spend much time to find what we need in many relevant web pages. It is more difficult to deal with the video information, but there is vast video information in the internet because of more and more people have become video creators. Therefore, the problem of how to recommend users relevant information has attracted many researchers\´ attention. Recommendation technology has developed quickly in the recent years. In this paper, we set up a video search engine with the open source project "Lucene", propose a semantic recommendation method and combine the search engine with the semantic recommendation technology. We do experiments to verify our method and analyze its shortcomings. From the experiments, we can see that the approach introduced in this paper can reach a better level than the traditional search engine. Both keywords and sentences can be processed by the system because the system understands them.