Method for dynamically optimizing expressive force evaluation questions

Abstract

The invention belongs to the technical field of statistical learning, is used for dynamically adjusting and automatically optimizing topic parameters in expressive force evaluation, and particularly relates to a method for dynamically optimizing expressive force evaluation topics, which applies a statistical learning method to improve the accuracy of evaluation judgment and dynamically adjust thedifficulty of the evaluation judgment while increasing the number of samples. The method comprises the following specific steps: 1) selecting a certain question qj={d, b, w, l}, wherein the d represents the difficulty of the question, the b represents the lexicon of the question, the w represents the weight parameter of the lexicon of the question, and the l represents the discrimination of the question; 2) optimizing a parameter d; and 3) optimizing the parameter w. According to the invention, a statistical learning method is used, useful information is extracted from continuously increased existing sample data to optimize topic parameters, so that a question model is more robust, and sensitive feedback is provided for different expressive force evaluation inputs; and by using the unsupervised learning model, extra labeling work is not needed in the learning process, the workload of experts is reduced, and the labor cost is reduced.

Description

technical field [0001] The invention belongs to the technical field of statistical learning, and specifically relates to a method for dynamically optimizing a test item of expressive power. Background technique [0002] The Rapid Expressive Test is a new oral test mode. Compared with the traditional Chinese language test, the Rapid Expressive Test has the advantages of wide application, diverse sources of questions, quick test and objective evaluation. However, at present, each attribute of the topic is stipulated by experts based on their own experience during initialization, and there may be inaccurate judgments. Aiming at this problem, the present invention provides a method, which can unsupervisedly adjust the question parameters by increasing the number of samples, and can improve the accuracy of question scoring to a large extent. Contents of the invention [0003] The purpose of the present invention is to provide a method for dynamically optimizing the test questi...

AI Technical Summary

Problems solved by technology

However, each attribute of the current topic is specified by experts based on their own experience during initialization, and there may be inaccurate judgments

Images