Research on Intelligent Question Answering algorithm based on machine learning

Abstract

Aiming at the student consultation faced by the enrollment of experimental classes, a small sampleQuestion Answering System was constructed, which can help students obtain enrollment status in time. First, I collected information and data about a university and experimental class, and designed a text similarity matching algorithm that combines word vectors and corpus tags to apply to the Question Answering System. Based on the word vectors of the neural network training corpus, the corpus is grouped using a clustering algorithm. The keyword extraction algorithm is used to extract the group tags. In the question sentence similarity calculation, the category similarity calculation is first performed, and then the sentence similarity calculation is performed in the category. In the case of a small sample, compared with the traditional sentence vector similarity calculation method, the accuracy rate in the Question Answering System is 64%, which meets the requirements of the Question Answering System in this field.