Research on automatic summary of Chinese short text based on LSTM and keywords correction
Fang Xu, Yi Guo, Qi Wang, Zhen Fan · 2018
Network platform is an important way for people to exchange information and communicate with each other. In order to ensure users to get the information they want faster, summaries should be generated from these short texts. However, such a large number of summaries have gone far beyond the limit of manual processing, so the automatic summary techniques need to be used to automatically generate the summary. Chinese automatic summary technology starts later, and the summary of short texts generated by the traditional method is not good. In this paper, deep learning method are introduced to generate short text summary automatically, and the keywords of the original text are also used to optimize the results. In the method, an attention-based sequence-to-sequence model with Long-Short Term Memory Network (LSTM) is constructed, which combines the character features and word features as its inputs. The keywords of short texts are used to correct the words of model-generated summary. Experiments show that our method improves the performance of short summaries generation on LCSTS dataset.