Improve question & answer system by applying genetic algorithm

In E-learning system, Q&A (question and answer) systems are important aiding tools for helping students to obtain information from the Internet. The systems find answers by matching the features of previous questions to current ones, and adjusting the weights of those features as users´ access to it. Since an existing Q&A engine maintains feature weights gradually according to the users´ response from the beginning of each query, the prediction accuracy is too limited, and the maintenance process is slow and inefficient. In order to improve prediction accuracy, We introduce genetic algorithm (GA) into the traditional Q&A system and put forward a new architecture for a Q&A engine which uses the conception of case based reasoning. The experimental results show that the prediction accuracy is greatly improved by our GA-based engine over other engines.