A Study on Improving Japanese Writing Skills by Constructing Japanese Syntactic Analysis and Generation Technology Using Computational Methods and Artificial Intelligence Models

In this paper, Japanese syntactic analysis and text embellishment techniques are designed to improve students' Japanese writing skills.Since Japanese dependency parsing is an important part of Japanese syntactic analysis.In this regard, this paper adopts the SVM model to generate a classifier using the labeled corpus as a way to determine whether there is a dependency relationship between two text sections.In order to improve the parsing accuracy of the SVM model, this paper proposes a Japanese dependency parsing method based on NN-LSVM pruning of a large-scale training corpus for dependency parsing on the basis of SVM.After that, a text touch-up technique based on syntactic structure is designed, which introduces a contrastive representation learning method and pushes the model to deeply understand the modeling relationship between semantic and syntactic structural information by adjusting the loss function to further mine more appropriate syntactic structures and expressions in order to improve the effect of the text touch-up technique.After verifying that the two techniques are feasible, this paper designs a practical task for teaching Japanese writing.During the practice, the Japanese writing scores of the experimental class using the NN-LSVM model and the language generation model designed in this paper for writing tutoring improved significantly (P<0.05), and there was no significant change in the control class.It shows that the technique in this paper can have the effect of promoting students' Japanese writing ability.