Study On A Selective Query Recommendation Method Based On System Concept Learning Support Ability

Search exploratory gives description to the users' behaviors when he becomes unfamiliar with the search field,or carries out the complex search tasks.In the exploratory search process,sometimes,the user found that he can't find the infomation he needs through the recommended queries given by other algorithms,and sometimes,he find that through those recommended queries,the information he wants appears to be quite less and even some infomation is of no use to his search,which often bring some awful experiences to the user himself,and some will lead to the user's discontinued search in this process.Thus,in the exploratory search process,there is the need for selection in the recommended queries given by other algorithms so that we can help users complete their exploratory searches successfully.Concerning the above-mentioned problems,this thesis studys the recommendation method of selective query that is based on the system-concept learning-support ability.First,when studying the system's supporting capacity,there are the following three problems need to be solved:(1)How to measure the system's supporting ability by the amount of resources,(2)How to measure the system's supporting ability by the availability of resources,(3)How to measure the system's supporting ability by the searching value of resources.Concerning the above-mentioned three questions,this thesis will study the evaluation for system's supporting ability from the following three areas:(1)evaluation for resources quantity-based system-supporting capacity,(2)evaluation for resources-use-value-based system-supporting capacity,(3)evaluation for resources-exploration-value based system supporting capacity.And besides,concerning the selective query recommendation,this these will study a comprehensive sorting and selecting approach,and with that,to bring into reality the selective-query recommendation that is based on the system-concept learning-support capacity.Firstly,for the realization of evaluation for resources quantity,this thesis puts forth an evaluation method that is based on information-entropy.Secondly,and concerning the evaluation for resources-use-value,in the thesis,the user log will be analyzed offline to build model for resources-use-value,and with this model to complete the evaluation for resources-use-value-based system-supporting capacity Then again,concerning the evaluation for resources-exploration-value,in the thesis,the user log will be analyzed to build the model for resources-exploration-value.With this model,thus completing the evaluation for resources-exploration-value based system supporting capacity.And finally,after the completion of the evaluation for system supporting capabilities,the sorting algorithm that based on the comprehensive weight is applied to complete the selective-query recommendation that is based on the system-concept learning-support capacity.In order to verify the result of computing and comprehensive sorting method for system supporting capacity,this thesis will conduct the experiments in measurement and analysis objectively and subjectively.And the experimental result shows that,the recommendation method of selective query that is based on the system-concept learning-support ability can better help the users carry out the exploratory searches.