implicating of Data Mining Techniques to Suggesting Queries in Information Retrieval Systems

Data mining detects patterns in the massive volume of data used in many disciplines. It can also be useful in our field, especially in information retrieval. In information retrieval, first the system-oriented paradigm and then the user-oriented paradigm have been introduced, the second paradigm being concerned with information needs. In the second paradigm, the inclusion of inappropriate queries is considered the main reason for not retrieving relevant documents. Therefore, one of the main topics of this paradigm is proposing and extending the appropriate query in the recommender system that can be used for data mining methods. There are four important methods to propose a query to strengthen the recommender system. The time series rule is one of these methods that deal with query frequency in a particular time unit. Another method is the association rule that addresses the dependency and association of queries. In addition to the dependence and association of queries, the order of query terms is also considered in the method of Association rule with Levenshtein distances. However, in all three of these methods, the log file is used, while in probabilistic theory, the document words are used to repair the lexical gap between the queries and the documents. Therefore, it seems that using probability theory to suggest the query yields better results.