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

Multi-attribute target classification method based on sample set

A classification method, a technique for sample sets, applied in the fields of instruments, character and pattern recognition, computer parts, etc.

Pending Publication Date: 2021-04-23
XI AN MIKESI INTELLIGENT TECH CO LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0020] The present invention provides a progressive, sample-based classification method for multi-attribute targets, especially for targets with clear attribute meanings, which solves the above problems of existing classification methods

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-attribute target classification method based on sample set
  • Multi-attribute target classification method based on sample set
  • Multi-attribute target classification method based on sample set

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0052] Example 1: Basic Classification Method

[0053]The process flow of the classification method of the multi-attribute target based on the sample set in the present invention is as follows: Figure 6 shown. by figure 1 Take the known sample set shown as an example. There are two types of samples in the known sample set, one is triangular and the other is five-star. Each sample has two attributes, A and B, and the category is a subset of five-star samples. The value range of the A attribute is [2,9], the value range of the B attribute is [5,9]; the value range of the A attribute of the sample subset whose category is triangle is [3,10], and the B attribute The value range of is [1,6]. From the perspective of the entire sample set, the value range of attribute A is [2,10], and the value span is 10-2=8; the value range of attribute B is [1,9], and the value span is 9-1 =8.

[0054] Targets to be classified such as figure 2 As shown in the circle, the purpose of classif...

Embodiment 2

[0058] Example 2: The strategy for the order of attribute selection in the sample set screening process

[0059] In the first embodiment, the difficulty of obtaining the values ​​of target attribute A and attribute B is not considered, so it is not important to filter from attribute A first or filter from attribute B first. However, the cost of obtaining different attributes of the actual target is different. For example, for a cup, the color of the cup is easy to obtain, while the density of the cup is more difficult to obtain. In the classification process, if we can directly implement classification from an easily obtained attribute, there is no need to obtain hard-to-obtain attributes and reduce the cost of classification.

[0060] That is to say, in the process of traversing attributes to filter sample sets, it is a more cost-effective classification method to preferentially filter from the threshold range of attributes with lower cost.

Embodiment 3

[0061] Example 3: The selected sample subset is empty, and the target to be classified is manually classified and added to the sample set

[0062] Such as Figure 5 As shown, the value of attribute A of the target to be classified is 12, and the value of attribute B is 11. The screening threshold range of attribute A is [11,13], and the screening threshold range of attribute B is [10,12]. From the figure It can be seen that no matter which attribute is used to filter, the obtained sample subset is empty.

[0063] At this time, the target to be classified can be submitted to manual classification. After the target category is manually determined, the target is added as a sample to the sample set as a sample set to be screened. If there are targets with similar values ​​that need to be classified later, you can pass the aforementioned classification. method to get the target category.

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-attribute target classification method based on a sample set, and the method comprises the steps: screening a sample subset according to a known sample set and the known attributes of a to-be-classified target, and determining that the type of the to-be-classified target is a single type if the screened sample has a single type attribute; otherwise, selecting a new known attribute of the to-be-classified target, and continuously screening the screened sample subset until the classification is completed or fails. According to the method, the interpretability of classification can be achieved, the improvement directions of the sample set and the classification method are given, a complex multi-dimensional classification problem is simplified into single-dimensional sample screening, an exploratory target classification method is provided, and the efficiency of the classification method is improved.

Description

technical field [0001] The invention belongs to the technical fields of machine learning and artificial intelligence, and in particular relates to a classification method. Background technique [0002] The goal of the classification problem is to determine which known sample class a new sample belongs to based on certain characteristics of known samples. According to the number of categories, classification problems can be further divided into binary classification and multiclass classification. [0003] For example, in email management, classifying an email as "spam" or "non-spam" is a typical binary classification problem; banks classify credit ratings of credit card customers and classify stock types of listed companies. Multivariate classification problems. [0004] After years of development, machine learning has developed numerous classification methods. Common classification methods include: [0005] linear classifier [0006] naive bayes classifier [0007] Perc...

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): G06K9/62
Inventor 邓少冬
Owner XI AN MIKESI INTELLIGENT TECH CO LTD
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