The tremendous growth in the amount of information available and the number of visitors to web sites in the recent years poses some key challenges for information filtering and retrieval. Web visitors not only expect high quality and relevant information, but also wish that the information be presented in an as efficient way as possible. The traditional filtering methods, however, only consider the relevant values of document. These conventional methods fail to consider the efficiency of documents retrieval. In this paper, we propose a new algorithm to calculate an index called document similarity score based on elements of the document. Using the index, document profile will be derived. Any documents with the similarity score above a given threshold will be clustered. Using these pre-clustered documents, information filtering and retrieval can be made more efficient.