This paper proposed a method for improving the XLNet model to address the shortcomings of segmentation algorithm for processing Chinese language. such as long sub-word lengths. long word lists and incomplete word list coverage. To address these issues. https://allfixelectricales.shop/product-category/vacuum-sound-proofing/