Question answering method facing specific field

A technology for specific fields and problems, applied in special data processing applications, instruments, electrical digital data processing, etc., can solve problems such as inaccurate identification of proper names

Inactive Publication Date: 2017-06-13
HARBIN INST OF TECH
View PDF3 Cites 37 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] The purpose of the present invention is to solve the problem that the existing technology is relatively accurate in identifying entities such as person nam

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
  • Question answering method facing specific field
  • Question answering method facing specific field
  • Question answering method facing specific field

Examples

Experimental program
Comparison scheme
Effect test

specific Embodiment approach 1

[0024] Specific implementation manner 1: The specific process of a specific domain-oriented question answering method of this implementation manner is:

[0025] Step 1: Build a vocabulary in a specific field, and use the vocabulary to segment the input question;

[0026] Step 2: Analyze the input problem after word segmentation, and identify the problem type and problem component;

[0027] Step 3: Expand the question components at the semantic and string level to obtain answer candidate words;

[0028] Step 4. Perform answer candidate word-attribute search in the knowledge base to obtain answer candidate paragraphs;

[0029] Step 5. Screen candidate answer sentences from answer candidate paragraphs.

specific Embodiment approach 2

[0030] Specific embodiment two: this embodiment is different from specific embodiment one in that: in the first step, a vocabulary of a specific field is constructed, and the input question is segmented using the vocabulary; the specific process is:

[0031] First crawl the title of the specific domain Baidu Encyclopedia, after deduplication operation, get the initial specific domain dictionary, use the initial specific domain dictionary to segment the specific domain data, get an initial segmentation result, and then use the initial segmentation result to train the segmentation device in the specific domain (through The C RF++ tool inputs the marked initial word segmentation results to obtain a word segmenter in a specific field). The specific field data is segmented using a specific field tokenizer, and then the word frequency of each unregistered word (the number of times the word appears in the specific field data) is extracted, and the word frequency is greater than a specifi...

specific Embodiment approach 3

[0041] Specific embodiment three: This embodiment is different from specific embodiments one or two in that: in the second step, the input problem after word segmentation is analyzed to identify the problem type and problem component; the specific process is:

[0042] Problem analysis includes problem classification and problem component labeling;

[0043] Question classification can not only guide the retrieval of answer candidate paragraphs, such as the comparison type requires the establishment of multiple queries, but also play an auxiliary role in the final answer generation.

[0044] The classification system of question classification is based on the answering mode given in the reference teaching aids; it is obvious that China’s college entrance examination has some fixed formulas, so classification according to the answering methods in the reference materials can improve the final Score. Some categories can be divided into subcategories, such as comparative problems, which c...

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

Provided is a question answering method facing a specific field. The invention relates to the question answering method facing the specific field. The goal is to solve a problem that in the prior art, identification of entities such as personal name, geographic name and organization name is accurate, but identification of proper names in the specific field is inaccurate is inaccurate. The process comprises a first step of constructing a word list in a specific field, and utilizing the word list to segment input questions; a second step of performing question analysis on the input questions having been divided; a third step of performing semantic questions and character string layer extension on question components, and obtaining answer candidate words; a fourth step of performing answer candidate word-attribute retrieval in a knowledge base, and obtaining answer candidate paragraphs; and a fifth step of screening candidate answer sentences from the answer candidate paragraphs. The question answering method is used for question answering in the specific field.

Description

Technical field [0001] The invention relates to a problem solving method oriented to a specific field. Background technique [0002] Question Answering System (QA) is an advanced form of information retrieval system, it can use accurate and concise natural language to answer users' questions in natural language. For us today, time is extremely precious, so it makes sense to build a question-and-answer system in a specific field. [0003] At present, there are not many related materials on the construction of question and answer systems in specific fields, but there are relevant information about question and answer systems based on structured data. The main idea of ​​a question and answer system based on structured data is to analyze the problem and transform it into a query (query) , And then query in structured data, and the returned query result is the answer to the question. The main data processing flow is as follows: (1) Analyze the problem according to the characteristics ...

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
IPC IPC(8): G06F17/30G06F17/27
CPCG06F16/243G06F16/90332G06F16/90344G06F40/216G06F40/284G06F40/30
Inventor 郑德权杨沐昀朱聪慧俞可李依尘赵铁军徐冰曹海龙
Owner HARBIN INST OF TECH
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products