Association Based Evaluating System For Entities Of Literature

How to evaluate quality of literatures so that people can retrieval high-quality literatures effectively from voluminous literatures becomes an urgent problem to solve. Most of current information retrieval systems have drawbacks, such as single criterion, limited objects evaluated, lacking of mechanisms for user's interaction in evaluation for literatures.To tackle with these problems, in Association based Evaluating System for Entities of Literature-AEEL, we present evaluating approach based on an associate network for entities of literature. The basic idea of the model is as follows. With the analysis of associate relations among entities of literature, we build an associate network for entities of literature, and adopt relevant algorithms to calculate various kinds of nodes dynamically and iteratively. Therefore, we can have the important evaluation for entities of literature. In consideration of the newly-published papers which are lack of referenced relations might result in that objective evaluation results can't reflect the quality of literatures, AEEL uses the subjective methods that are based on the users'feedbacks to make up for the inaccuracy. Subjective evaluation records and analysing users' behaviour quantitativly in retrieval system are criterions of evaluating for literatures, which can cover the drawback with just using associate relation and reflect popularity of lteratures timely. The objective method of literature evaluation is from the view of the associate relations among the entities of literature, whereas the subjective method is based on the users'feedbacks. To combine advantage of these two different approaches of evaluation,we choose the comprehensive evaluation which combines objective evaluation and subjective evaluation. The essence of comprehensive evaluation is As follows. Taking samples from the literatures with sufficient associate relations, finding out the internal relations between the two methods. According to the relations, the evaluation can assign weights for the two approaches mentioned before to make more accurate evaluating results.Based on the Windows operating system platform, association based Evaluating System for Entities of Literature is implemented by using Java language. Test is done based on this system. It's concluded that associate network based evaluating approach for entities of literature is more objective and accurate than other link based evaluating approach through analying and comparing associate network based evaluating approach for entities of literature with other dynamic iterative methods and static statistical methods. The system is applied in retrieval platform, and it's normal to return results according to evaluating results of entities of literature. Users can evaluate literature by employing a variety of methods, the interaction betweent users and system is strong.