Study on a speech learning approach based on interval support vector regression

In this paper an interval regression model has been established to cope with the situation that the input training data is accurate while the output one is interval. And the model has been applied in English speech learning system to predict the credible interval of correct speech and then give a correct judgment for the learners. Experimental data show that the new model reduces the workload of fuzzy prediction and has good accuracy, so it can be effective in speech learning system.