Segmented sentence recognition method and device for human-machine intelligent question answer system

Abstract

Disclosed are a segmented sentence recognition method and device for a human-machine intelligent question-answer system. One particular embodiment of the method includes: receiving a current sentence input by a user; inputting the current sentence into a pre-trained segmented sentence recognition model to obtain a first probability of the current sentence being a complete sentence, wherein the segmented sentence recognition model is used for determining the first probability according to a probability corresponding to a Chinese language model of the current sentence and a probability corresponding to a recurrent neural network language model; and if the first probability is greater than a pre-set first threshold value, determining the current sentence to be a complete sentence. In the embodiment, a Chinese language model and a recurrent neural network language model are merged to obtain a segmented sentence recognition model for determining a probability of a current sentence being a complete sentence.