An effective LDA-based time topic model to improve blog search performance

Blog search engines and general search engines automatically crawl web pages from the Internet and produce search results for users. One difference between the two is that blog search engines focus on posts and ignore the rest of the pages. Obviously, the pages indexed by the general search engine are always greater than the posts. This feature allows bloggers to focus only on the posts they are interested in, rather than other types of pages. The other difference is that posts involve more time-related issues compared to general pages. For the general pages, the general search engine often can only show the last update time. However, for the post, the blog search engine can display various possible times. For some often updated posts, the time factor can help bloggers find information more efficiently. In this paper, we first use some well-known semantic analysis models to analyze the performance of the blog search. Next, we consider the time relationship between posts to further improve its performance. Finally, we provide some experiments to simulate various possible scenarios to confirm the effectiveness of this relationship. The contributions of this paper are twofold. One is that we build a high-performance system that considers the importance of blog topics at different times. The other is that we consider the time relationship between posts, which can rank the relevant blog topics based on the popularity of the posts. (C) 2017 Elsevier Ltd. All rights reserved.