A multi-label intelligent marking method and system

Abstract

The invention discloses a multi-label intelligent marking method and system, which belong to the field of computer technology, and obtain m candidate standard words to be matched associated with any label by searching in a self-built standard lexicon for preliminary screening; A label and m candidate standard words to be matched are calculated one by one to obtain the similarity between any label and each candidate standard word to be matched; the similarity threshold is set, and according to the similarity threshold, the candidate standard words to be matched Determine the n standard words to be matched related to any label. The multi-label intelligent marking method and system realizes as many standard words as possible in the self-built standard lexicon by manually constructing a label dictionary to find corresponding labels to realize large-scale multi-label technology, and improve the recall rate under the premise of ensuring accuracy; The flexibility is high, and tags can be added at any time; using elasticsearch to search for preliminary screening combined with similarity calculations, the standard words that match any tag can be finally obtained, which can meet the needs of high-concurrency commercial systems and improve the recall rate.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a multi-label intelligent marking method and system. Background technique [0002] In the application process of medical data, we need to perform multi-label marking tasks on the words in the standard lexicon. The traditional manual-based method is inefficient and can no longer meet normal production needs. Large-scale multi-label learning has been widely used in practical applications such as document annotation, search ranking and product recommendation, and is an important research problem in current computer technology. [0003] An existing large-scale multi-label learning technique is to construct a classifier, which can automatically select a subset of labels most relevant to standard words from a very large-scale label collection to mark standard words. The difficulty of this type of technology is that both the data feature space and the label space have extremely high...

AI Technical Summary

Problems solved by technology

The difficulty of this type of technology is that both the data feature space and the label space have extremely high dimensions and are sparse, requiring a large amount of manual data labeling, which is time-consuming and labor-intensive.

For example: Wikipedia's label dimension L is millions, then there will be 2L possible label subsets, the data is huge, and the traditional labeling method is obviously no longer applicable

Images