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

A classification and prediction method based on multi-stage hybrid model

A hybrid model and classification prediction technology, applied in genetic models, genetic rules, character and pattern recognition, etc., can solve problems such as long calculation time, ineffective research and exploration, and high computational complexity

Inactive Publication Date: 2019-01-18
ZHEJIANG UNIV OF FINANCE & ECONOMICS
View PDF0 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, how to choose the most effective ensemble model for different datasets has not been effectively studied and explored until now.
In addition, since the computational complexity of classifier selection in raw data is usually high and the computation time required is very long, it is necessary to explore a more efficient classifier selection method to obtain a suitable ensemble model that controls the complexity within acceptable range

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 classification and prediction method based on multi-stage hybrid model
  • A classification and prediction method based on multi-stage hybrid model
  • A classification and prediction method based on multi-stage hybrid model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0049] Embodiment 1, common multi-population niche genetic algorithm is exactly to add niche operation on multi-population genetic algorithm, concrete steps are as follows:

[0050] Step S1.1, initialization, generating a preset number of initial populations, and using the generated initial populations as the current population;

[0051] Step S1.2. For each current population, evaluate the population according to the fitness function corresponding to the candidate classifier;

[0052] Step S1.3, judging whether the termination condition of the iteration is satisfied, if it is satisfied, the iteration is ended and the optimal individual is output, otherwise, enter the next step;

[0053] Step S1.4, performing a selection operation on the current population;

[0054] Step S1.5, performing a cross operation on the current population;

[0055] Step S1.6, performing a mutation operation on the current population;

[0056] Step S1.7, perform niche operation on the current populatio...

Embodiment 2

[0059] Embodiment 2. Improved Multi-population Niche Genetic Algorithm. This embodiment combines multiple filtering methods to determine the importance of comprehensive features of all features. Based on the importance of comprehensive features, the original features are reordered, and according to their comprehensive Feature importance deletes some features in advance, and then uses the multi-population niche genetic algorithm to obtain the optimal feature subsets corresponding to different classifiers. The specific process is as figure 2 shown, including the following steps:

[0060] Step S2.1, using the hybrid filtering method to calculate the comprehensive feature importance of each feature, filter out the features whose comprehensive feature importance is not less than the set threshold, calculate the probability of the selected feature being selected according to the comprehensive feature importance, and generate a preset The number of initial populations, with the ini...

Embodiment 3

[0119] Embodiment 3, in the classifier selection stage, the multipopulation niche genetic algorithm adopted comprises the following steps:

[0120] Step 3.1. Obtain the probability of each candidate classifier being selected based on the predictive prior knowledge of the candidate classifiers, generate a preset number of initial populations, and use the initial population as the current population;

[0121] Step 3.2, use the corresponding fitness function to evaluate the population for each current population;

[0122] Step 3.3, judging whether the iteration termination condition is satisfied, if it is satisfied, then end the iteration and output the optimal individual, otherwise enter the next step;

[0123] Step 3.4, select the current population;

[0124] Step 3.5, perform cross operation on the current population;

[0125] Step 3.6, performing a mutation operation on the current population;

[0126] Step 3.7, carry out niche operation on the current population;

[0127...

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 classification prediction method based on a multi-stage hybrid model, which adopts a multi-population niche genetic algorithm and combines a plurality of filtering methods and classifier prediction prior knowledge respectively in the process of feature selection and classifier selection, thereby obtaining an optimal feature subset and an optimal classifier subset. Then, the classifier ensemble method is used to integrate the optimal classifier subset with the optimal feature subset into the overall model for the final prediction. Finally, the hybrid model is applied to the field of credit rating to verify its forecasting performance in binary classification problem. The experimental results show that the multi-stage method used in the hybrid model plays a positiverole in improving the prediction performance of the model, and the final prediction performance of the model is better than that of other comparative models.

Description

technical field [0001] The invention belongs to the technical field of classification prediction, in particular to a classification prediction method based on a multi-stage mixed model. Background technique [0002] In recent years, the research and application of artificial intelligence and machine learning technology have made remarkable progress. In order to improve the predictive performance of binary classification, people started from various aspects and established a variety of new models. Among them, the credit evaluation model is a typical application of artificial intelligence and machine learning technology in binary classification prediction. Because of its important role in credit risk management, credit evaluation has received extensive attention from the financial industry. A small improvement in credit evaluation models can bring huge benefits to financial institutions. To this end, artificial intelligence and machine learning models have been applied to c...

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): G06K9/62G06N3/12
CPCG06N3/126G06F18/2113G06F18/24
Inventor 张文宇张帅何红亮裘一蕾
Owner ZHEJIANG UNIV OF FINANCE & ECONOMICS
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