Merging Strategy Based on Logistic Regression

With the development of network technology, the users looking for information may send a request to various selected databases and then inspect multiple result lists. To overcome such multiple inspections, the database merging strategy involves the merging of the retrieval results produced by separate, autonomous servers into an effective, single ranked list. To achieve this merging, this study is concerned with a particular situation within which only the rank of the retrieved records is available as a key to combine different result lists. Based on this rather limited information, this paper describes the theoretical foundation and retrieval performance of our database merging approach based on the logistic regression.