Patent and thesis joining method and system, and storage medium

The invention belongs to the technical field of information processing, and particularly relates to a patent and thesis joining method and system and a storage medium. The connection method comprises the following steps: S1, performing translation deduplication, preprocessing and word segmentation processing on patent literatures to obtain corresponding patent word sequences; s2, label sequences corresponding to the patent word sequences are manually marked, a first training set is formed by the multiple patent word sequences and the corresponding label sequences, and after preliminary training is conducted on the basis of the first training set through a multi-label classification model, the label sequences of the patent word sequences are formally output through the multi-label classification model; s3, using a CRF model to output an optimal tag sequence corresponding to the current tag sequence based on the tag sequence of the current patent document; and S4, after the labels in the optimal label sequence