Unlock instant, AI-driven research and patent intelligence for your innovation.

A method and system for identifying factuality of Chinese events

A recognition method and factual technology, which can be used in text database indexing, unstructured text data retrieval, text database clustering/classification, etc., and can solve problems such as low accuracy, inapplicability, and low recall.

Active Publication Date: 2020-05-15
SUZHOU UNIV
View PDF3 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] At present, there are three problems in the main Chinese event de facto analysis methods: 1) The factuality of events is analyzed by formulating rules, and the efficiency of recognition depends largely on the quality of the rules, so experts and scholars in related fields are needed formulate
This requires higher cost and does not have universal applicability
2) Under the current rule method, the imbalance of categories leads to a more serious imbalance in recognition performance. Categories with a large number of events can achieve a good recall rate, but the accuracy rate is not high, and those with a small number of events Category can achieve good precision, but low recall

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
  • A method and system for identifying factuality of Chinese events
  • A method and system for identifying factuality of Chinese events
  • A method and system for identifying factuality of Chinese events

Examples

Experimental program
Comparison scheme
Effect test

example 2

[0175] Example 2: Modality: Other / Tense: Unspecified / Source: Prosecutors and Police / ESP_Word: Worry / ESP_Level: Possible / Degree_Word: Possible / Degree_Level: Possible / Degree_Tense=none / Negative: No / Facutuality: May not happen.

example 3

[0176] Example 3: Modality: Other / Tense: Unspecified / Source: Prosecutors and Police / ESP_Word: Worry / ESP_Level: Possible / Degree_Word: Possible / Degree_Level: Possible / Degree_Tense=None / Negative: No.

[0177] S20, on the tagged corpus, for the factual related information of each Chinese event, use the method of rules to process, transform and fuse the features, obtain a series of factual related features, and then add the real factuality of the event, and then Construct a collection of labeled corpus features;

[0178] On the test corpus, for the factual related information of each Chinese event, use the same rule method to process, transform and fuse the features to obtain a series of factual related features, and then construct the test corpus feature set.

[0179] Among them, such as image 3 As shown, the specific process of S20 is as follows:

[0180] S201, event sentence feature processing, selecting the event sentence modality and temporal information to which each event...

example 4

[0184] Example 4: .

[0185] S202, lexical-level feature processing, performing part-of-speech tagging on the event source, negative words and degree words of each event, and then selecting these three parts-of-speech as lexical-level features, and adding them to the corpus feature set.

[0186] In the tagged corpus, use the part-of-speech tagging tool to tag the three types of lexical information, event source, negative word, and degree word, and select their part-of-speech as a feature. If the current event does not have one of the above words, its corresponding part-of-speech feature will be defaulted is "none", and these three types of information are added to the corpus feature set.

[0187] In the test corpus, use the part-of-speech tagging tool to tag the three types of lexical information, event source, negative word, and degree word, and select their part-of-speech as a feature. If the current event does not have one of the above words, its corresponding part-of-speec...

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 relates to a Chinese event factuality recognition method and system. Factuality of a Chinese event is recognized by adopting the method of combining machine learning and inference and utilizing event factuality information and relationships among the information. Compared with a current method and system, the Chinese event factuality recognition method and system are improved in overall recognition performance, meanwhile have a better effect in treating the unbalance problem of categories, and have obvious performance improvement in factuality recognition of the categories with a small number of the events.

Description

technical field [0001] The invention relates to the field of natural language processing, in particular to a method and system for identifying the factuality of Chinese events. Background technique [0002] When people talk about an event and express their views and thoughts on the event, they not only convey information such as the time, place, and person of the event, but also include their position and attitude towards the event. Among them, attitudes and positions can be divided into two categories: subjectivity and certainty. Subjectivity is the narrator's opinion on the subjectivity of current events, such as agreeing, opposing or neutral. [2] . Certainty refers to the narrator's degree of certainty about whether the current event is true or not, such as it must happen, it may happen, or it has not yet happened, etc. [3] . The certainty here refers to the factuality of events referred to in this article. [0003] Event factual recognition is to determine the degree...

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): G06F16/31G06F16/35G06F16/36
CPCG06F16/313G06F16/355G06F16/36
Inventor 何天雄李培峰朱晓旭朱巧明周国栋
Owner SUZHOU UNIV