Method and Research on Automatic Generation and Fluency Discrimination of Movie Word Sequence Based on NNLM

In the context of massive information on the Internet, how to automatically generate word sequence statements based on words and make them compliant is crucial. In this paper, the neural network language model (NNLM) is used to judge the automatically generated Chinese movie title sentence, and the highest probability value in the data set is obtained, that is, the top 10 statements with the best fluency. At the same time, the N-Gram algorithm is used for screening, and the screening results are compared with the results of NNLM. After professional evaluation and analysis, the discriminant result of NNLM algorithm is better. However, in the real-time discriminating scene, the computational complexity of the NNLM algorithm is much larger than that of the N-Gram algorithm, so it needs to be considered after comprehensive consideration.