English text quality analysis based on recurrent neural network and semantic segmentation

Abstract

In recent years, deep learning algorithm based on cyclic neural network and semantic segmentation has performed well in the field of image segmentation. The purpose of this paper is to realize the quality analysis of English text through recurrent neural network and semantic segmentation. This paper proposes an attention based English text quality analysis model based on recurrent neural network. Through the introduction of attention mechanism, the influence of semantics in the text is considered in the analysis of English text quality. The target relies on the quality of English text to determine the text quality of the sentence for a given target object. At present, most English text quality analysis methods are aimed at the traditional semantic analysis tasks. Based on rnn-attention model, a rnn-attention-t model is proposed, which introduces the information of the target object while modeling the text. In addition, considering that the influence of the top and bottom of the target object on the semantic trend is usually different, this paper proposes an rnn-attention-c model, which models the top and bottom of the target object respectively. The experimental data have shown that the quality analysis of English text based on recurrent neural network and semantic segmentation is faster than the traditional method. The experimental results have demonstrated that our method can effectively and quickly confirm the quality of English text, which is about 7% faster than the conventional method.