Code recommendation method and system, corresponding equipment and storage medium

Abstract

The invention discloses a code recommendation method and system, corresponding equipment and a storage medium. The method comprises the steps: obtaining a medical record home page diagnosis text as the input of a trained first sub-eye inference model, the diagnosis text at least comprises a main diagnosis text, wherein the first sub-eye inference model comprises a deep learning language model anda sub-eye projection layer linked with the output of the deep learning language model, and the deep learning language model is used for at least outputting a main diagnosis semantic vector based on aninput diagnosis text; in response to the fact that the diagnosis text does not include the secondary diagnosis text, the sub-view projection layer is used for determining the sub-view classificationprobability of the corresponding primary diagnosis text for each international disease classification ICD based on the primary diagnosis semantic vector; and taking the m sub-items with the highest classification probability for the main diagnosis text as recommended main diagnosis ICD sub-items of the main diagnosis text. According to the invention, corresponding ICD codes can be intelligently and automatically recommended based on texts, and the accuracy and the efficiency are high.

Description

technical field [0001] The present application relates to the field of electrical digital data processing, in particular to a coding recommendation method, system, corresponding equipment and storage medium. Background technique [0002] Traditional auxiliary coding tools have existed for many years, but the way to achieve intelligence is still at a relatively shallow level. Most of the existing coding tools are based on text, and provide possible hints based on keywords, and guide the coders to obtain the final code step by step. The core of traditional auxiliary coding is string search and string matching. [0003] However, most of the time the text is presented in natural language (i.e. manually entered textual content). At the same time, the standard writing method of nationally standardized ICD codes (including ICD-10 and ICD-9-CM3 versions) is also based on natural language. In general, you need to select the appropriate corresponding code from the ICD code automati...

AI Technical Summary

Problems solved by technology

However, in many cases, the writing method is completely different from the form in the national standard, so in some cases the corresponding ICD code cannot be found through keyword searches

Images