Abstract：Improving Diversity of Search Results for the National Library of Turkey One of the main objectives of search systems is to improve the quality of the user experience by providing search results that are relevant to the users information needProviding a diverse set of search results contribute to the quality of service of a search system as it increases the possibility of satisfying the users information needAlthough literature suggests that diversification techniques are used for the Web searchthey are not applied often to non-free text search systemsIn this workwe propose a diversified search framework for library catalog search systemsThe work presented here is tested on a real-world dataset acquired from the National Library of TurkeyWe have experimented the effects of changes in a wide range of parameters for our frameworktesting the effectivenessboth in terms of search result diversity and result relevanceTo this endthe proposed models adopt an automatic document categorization technique as a generic approach to the category discovery of the documents based on Latent Dirichlet AllocationLDAFor diversification of the search resultswe modified and used a Maximum Marginal RelevanceMMR-based techniqueThe proposed framework achieved significant improvements on search result diversity667while maintaining acceptable levels of relevance19lossCollection ClassificationIt is mentioned that the use of LDA for automatic documentation is mentionedwhich may be related to the collection classificationRetrieval and IndexingBecause this research involves the diverse search framework and the diversity of search resultsit is related to retrievalIn this studythe author uses the potential Dirichlet allocationLDAtechnology as a machine-based automatic document classification method to find the type of fileIn additionthey also modified and used the technologies based on the largest edge correlationMMRto diversify the search results