Recognition method and device

A recognition method and object technology, applied in the field of information recognition, can solve problems such as overfitting of a single model, affecting the accuracy of identifying invalid comments, etc.

Active Publication Date: 2017-09-22
TENCENT TECH (SHENZHEN) CO LTD
View PDF9 Cites 12 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, the Bayesian method is a common method for identifying invalid reviews. However, because this method assumes that there is no relationship (independent) between words and words in the review information, it does not match the reality. In text classification, the training samples are biased. The high-latitude characteristics of text features in natural language processing lead to overfitting of a single model after training, which will affect the accuracy of identifying invalid comments

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
  • Recognition method and device
  • Recognition method and device
  • Recognition method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0081] like figure 1 As shown, this embodiment provides an identification method, which includes:

[0082] Step S110: Obtain information objects to be classified; wherein, the information objects include at least user comment information.

[0083] Optionally, the information object generally refers to a text object.

[0084] For example, the information object may be a comment sent by a user on a certain article or a certain Weibo. Here, the format type of the review content is not limited, and the review content may be text, or pictures, or audio, or video.

[0085] Step S120: Use the trained M different classification models to classify the information objects to be classified respectively; wherein, M is a positive integer greater than or equal to 2.

[0086] The M different classification models described in this embodiment are M different classification models, and this difference may be reflected in different types of classification models, or different training data o...

Embodiment 2

[0174] like Image 6 As shown, this embodiment provides an identification device, which includes:

[0175] An acquisition unit 61, configured to acquire information objects to be classified; wherein, the information objects include at least user comment information;

[0176] A classification unit 62, configured to use the trained M different classification models to classify the information objects to be classified respectively; wherein, M is a positive integer greater than or equal to 2;

[0177] A statistical unit 63, configured to count the first number information of classifying the information object to be classified into the first type of information object in the M classification models, and classify the information object to be classified into the second type The second number information of the information object;

[0178] A determining unit 64, configured to determine a final classification of the information object to be classified based on the first number inform...

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 discloses a recognition method an device. The recognition method comprises the steps that a to-be-classified information object is acquired, wherein the information object at least comprises user comment information; trained M different classification models are utilized to classify the to-be-classified information object, wherein M is a positive integer greater than or equal to 2; first number information obtained by classifying the to-be-classified information object into a first-class information object and second number information obtained by classifying the to-be-classified information object into a second-class information object in the M classification models are obtained through statistics; and a final class of the to-be-classified information object is determined based on the first number information and the second number information.

Description

technical field [0001] The invention relates to information identification technology, in particular to an identification method and device. Background technique [0002] With the advent of the Internet age, people's speech on major websites or forums is more free and casual. Major websites or forums receive a lot of user comment information every day. For example, there may be a large number of invalid comments such as advertisements, abuse, and pornographic information in some product comments. Therefore, how to manage comment information will face a severe test . [0003] At present, the Bayesian method is a common method for identifying invalid reviews. However, because this method assumes that there is no relationship (independent) between words and words in the review information, it does not match the reality. In text classification, the training samples are biased. The high-latitude characteristics of text features in natural language processing lead to overfitting...

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): G06F17/30G06K9/62
CPCG06F16/35G06F18/241
Inventor 黄鹏
Owner TENCENT TECH (SHENZHEN) CO LTD
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

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap