Extracting Structured House Information from BBS Based on Prolog

BBS provides users with a space of free communication and plentiful information resources. However, to gain manually useful information from constantly updated, huge and unstructured data is very difficult for users. This paper applies Prolog to BBS data mining, and builds a housing information mining system based on Prolog, which extracts structured house leasing information from the large number of unstructured text in BBS. Firstly, we develop a BBS crawler, which obtains relevant documents from specific BBS sites and stores them to local directory. Then, by analyzing all crawled unstructured documents, we extract useful information (e.g. building name, type and price etc.) and store them to a relational database, on which an interface provided for users to query. We experiment on a real-world BBS, the result shows that this Prolog programs can be applied to BBS data mining effectively.