Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Multi-classification intelligent question and answer retrieval method for medical custom entity word part-of-speech tags

A technology of intelligent question answering and self-defined words, applied in digital data information retrieval, unstructured text data retrieval, text database clustering/classification, etc. , to achieve the effect of fast data retrieval, improved accuracy, and improved speed

Inactive Publication Date: 2019-11-19
未来火种(北京)科技有限公司
View PDF4 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, natural language processing technology, as one of the important branches of the artificial intelligence application field, has been widely used in search, advertising and other fields. Due to the characteristics of the medical field, in the intelligent question and answer process of medical products, the returned answers Affect the patient's judgment, and inaccurate and non-authoritative answers may cause double losses of the patient's life and property

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Multi-classification intelligent question and answer retrieval method for medical custom entity word part-of-speech tags
  • Multi-classification intelligent question and answer retrieval method for medical custom entity word part-of-speech tags

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0046] Embodiment 1, self-defining medical entity word part-of-speech tag and writing retrieval template in the first part, concrete steps are as follows:

[0047] (1), according to the medical entities and relationships stored in the graph data, perform part-of-speech tags on the named entity words;

[0048] (2) For the manually processed classified user intent training data, replace named entity words with custom part-of-speech tags;

[0049] (3) Based on the user intention training data, write template statements for retrieving entities and relationships in the graph database.

[0050] The image data in step (1) uses Neo4j, and the part-of-speech tags of named entity words in step (1) cannot be repeated with the part-of-speech tags in the keyword thesaurus.

Embodiment 2

[0051] Embodiment 2, with reference to figure 1 , in the second part, the multi-classification model identifies user intent, and the specific steps are as follows;

[0052] (a), using the custom part-of-speech tag data replaced in step (2), determine the model training set sample;

[0053] (b), add custom words, named entity word part-of-speech tags;

[0054] (c), based on the medical stop word thesaurus, the medical keyword thesaurus, adding a custom word thesaurus, using natural language processing technology for word segmentation and keyword extraction, and building a vocabulary;

[0055] (d), based on the vocabulary constructed in step (c), the original text training data is carried out to text vectorization;

[0056] (e), using a multi-classification algorithm model, training the tagged word vectorization data, and constructing a multi-classification model;

[0057] (f), based on the multi-category model constructed in step (e), realize precise division of newly added ...

Embodiment 3

[0059] Embodiment 3, in the third part, the named entity part-of-speech tag replaces the real value, and the specific steps are as follows:

[0060] (1), in the second part, determine the user intent, and associate the user intent retrieval sentence in the first part;

[0061] (2), in the retrieval sentence, replace the part-of-speech tag of the named entity word with a real entity word;

[0062] (3) Connect the graph database to perform entity and relationship retrieval;

[0063] (4) Feedback the retrieval results to the questioner.

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

PUM

No PUM Login to View More

Abstract

The invention discloses a multi-classification intelligent question and answer retrieval method for medical custom entity word part-of-speech tags. The method comprises a first part, a second part anda third part, the first part is a custom medical entity word part-of-speech label and a compiling retrieval template, the second part is a multi-classification model identification user intention, and the third part is a named entity word part-of-speech label replacement true value. According to the multi-classification intelligent question and answer retrieval method for the part-of-speech tagsof the medical custom entity words, the medical knowledge graph is constructed and stored in the graph data neo4j, so that the incidence relation of the medical data is perfected, and the data retrieval speed is increased; the part-of-speech of the medical custom entity words is labeled, and meanwhile, the classified user intention training data entity words are replaced with custom part-of-speechlabels, so that the accuracy of predicting the user intention by the model is improved.

Description

technical field [0001] The invention relates to the field of intelligent question answering systems, in particular to a multi-category intelligent question answer retrieval method for medical self-defined entity word part-of-speech tags. Background technique [0002] The intelligent question answering system relies on natural language processing technology to analyze the real intention of the user's question, and returns the correct matching answer according to the matching degree of the candidate question answer. The intelligent question answering system mainly consists of user question analysis, information retrieval and answer generation; [0003] At present, natural language processing technology, as one of the important branches of the artificial intelligence application field, has been widely used in search, advertising and other fields. Due to the characteristics of the medical field, in the intelligent question and answer process of medical products, the returned answ...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More

Application Information

Patent Timeline
no application Login to View More
Patent Type & Authority Applications(China)
IPC IPC(8): G06F16/332G06F16/35G06F17/27
CPCG06F16/3329G06F16/35
Inventor 赵海艳
Owner 未来火种(北京)科技有限公司
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com