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

Faq similarity calculation method, device and electronic equipment based on knowledge enhancement processing

A similarity calculation and enhanced processing technology, applied in the field of similarity calculation, can solve problems such as insufficient semantics, limited amount of target text data, and large gaps

Active Publication Date: 2021-01-19
HUNDSUN TECH
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The amount of user questions and target text data in a single sentence is very limited, which makes the semantics of the two texts not rich enough, and it is easy to cause the result that the user question does not match the target text.
If the gap between the customer's question "How much is my handling fee" and the target text "A-share transaction fee" is too large, the direct calculation of the similarity may not match the correct answer

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
  • Faq similarity calculation method, device and electronic equipment based on knowledge enhancement processing
  • Faq similarity calculation method, device and electronic equipment based on knowledge enhancement processing
  • Faq similarity calculation method, device and electronic equipment based on knowledge enhancement processing

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0068] At the word level, collect several synonymous sentences of the target text as knowledge accumulation, segment the target text and its synonymous sentences into words, and form a corresponding knowledge distribution with the results after word segmentation; when calculating the similarity, use the knowledge distribution to compare the target text for knowledge enhancement.

[0069] In order to achieve the above technical purpose, according to the first aspect of the embodiment of the present disclosure, this embodiment provides a method for calculating FAQ similarity based on knowledge enhancement processing, such as figure 1 As shown, the FAQ similarity calculation method includes:

[0070] 11. Receive customer texts and determine the target text corresponding to the customer texts;

[0071] 12. Carry out knowledge enhancement processing based on knowledge distribution on customer text and target text, and perform FAQ similarity calculation based on the processing resu...

example 2

[0114] Customer question: Excuse me, what is the stamp duty for buying and selling stocks now, thank you;

[0115] Correct target text: A share transaction fee,;

[0116] Wrong target text: Shanghai B-share transaction fee;

[0117] A synonym for the correct target text: How much is my transaction fee. How much is the stamp duty on my stock purchases. Stock selling costs. How much is the transaction fee for A shares?

[0118] A synonym for the wrong target text: Shanghai B stock transaction fee. How much is the transaction fee for B shares in Shanghai? Excuse me, how much is the stamp duty for buying and selling B shares now, thank you.

[0119] Step 1: If the customer text is a short text, word segmentation is performed on the customer text and the target text corresponding to the customer text.

[0120] Customer questions: Excuse me, now, stock, buy, sell, of, stamp duty, yes, how much, thank you

[0121] Correct target text: A shares, transactions, fees;

[0122] E...

Embodiment 2

[0143] On the basis of the aforementioned embodiment one, if the obtained FAQ similarity result and the preset threshold are judged, if the FAQ similarity result is not greater than the preset threshold, generate synonymous sentences for expanding the customer text and the target text semantic range Sentence group, based on synonymous sentence group to calculate FAQ similarity. Such as figure 2 shown, including:

[0144] 311. Construct a synonymous sentence generation model.

[0145] The details of building a synonymous sentence generation model include:

[0146] 3111. Construct an initial deep learning model;

[0147] 3112. In the initial deep learning model, the long-short-term memory artificial neural network and the multi-layer perceptron are sequentially called to encode the received customer text into encoded vectors;

[0148] 3113. In the initial deep learning model, the multi-layer perceptron and the long-short-term memory artificial neural network are called agai...

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 embodiment of the invention provides a knowledge enhancement processing-based FAQ similarity calculation method and device, and an electronic apparatus. The method comprises the steps of receivinga client text, determining a target text corresponding to the client text, performing knowledge enhancement processing based on knowledge distribution on the client text and the target text, performing FAQ similarity calculation based on a processing result, and performing different processing according to a relationship between athe similarity calculation result and the preset threshold; judgingwhether to continue the operation of expanding the synonymous sentence groups of the semantic ranges of the client text and the target text or not based on the relationship between the FAQ similarityresult and the preset threshold value after the processing, and further performing FAQ similarity calculation based on the synonymous sentence groups. The target text is expanded in different modes in the two judgment results, the same word range of user questions and the target text is expanded, and the similarity calculation result is improved.

Description

technical field [0001] The invention belongs to the field of similarity calculation, in particular to an FAQ similarity calculation method, device and electronic equipment based on knowledge enhancement processing. Background technique [0002] FAQ (Frequently Asked Questions) similarity calculation is widely used in various business fields. For example, in the securities industry, major securities companies will summarize some questions frequently asked by customers, and summarize these questions into a series of target texts. Then use the FAQ similarity calculation method to select the answer corresponding to the text with the highest similarity from the target text as the best answer to the customer's question. There are many existing FAQ similarity calculation methods, which can be summarized into three categories: methods based on word matching, methods based on word matching and word vectors, and methods based on word vectors or sentence vectors. [0003] The above s...

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 Patents(China)
IPC IPC(8): G06F16/332G06F40/30G06F40/247
CPCG06F16/3329G06F40/247G06F40/30
Inventor 廖智霖高峰吕杨苗林金曙
Owner HUNDSUN TECH
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