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

Model training method and device, electronic equipment

A technology of model training and training samples, applied in the computer field, can solve problems such as inaccurate prediction results, achieve the effect of improving the accuracy of prediction results and avoiding uneven distribution of training data

Inactive Publication Date: 2018-07-13
BEIJING SANKUAI ONLINE TECH CO LTD
View PDF0 Cites 33 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] This application provides a model training method to solve the problem of inaccurate prediction results when the model trained by the model training method in the prior art is applied to applications such as data mining or search

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
  • Model training method and device, electronic equipment
  • Model training method and device, electronic equipment
  • Model training method and device, electronic equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0026] A model training method disclosed in this embodiment, such as figure 1 As shown, the method includes: step 110 to step 140.

[0027] Step 110, according to the training samples, obtain at least a subset of training samples.

[0028] The training samples used to train the model usually include: sample labels and sample features of preset dimensions. The preset dimensions are selected according to different application scenarios of the source data and the model to be trained. Taking the prediction of the user's purchase rate as an example, the possible preset dimensions include: user gender, age, occupation, place of residence, commodity category, price, number of purchases and other characteristics. The larger the number of training samples, the more accurate the prediction results of the trained model will be. During the specific implementation of the present application, the training samples are first sampled to obtain multiple sub-training samples, which are respec...

Embodiment 2

[0038] A model training method disclosed in this embodiment, such as figure 2 As shown, the method includes: Step 210 to Step 270.

[0039] Step 210: Obtain at least a subset of training samples according to the training samples.

[0040]The larger the number of training samples, the more accurate the prediction results of the trained model will be. During the specific implementation of the present application, the training samples are first sampled to obtain multiple sub-training samples, which are respectively used to train different machine learning models. During specific implementation, the obtaining at least one sub-training sample according to the training sample includes: performing random sampling on the training sample to obtain at least one sub-training sample; and performing feature sampling on each sub-training sample. Assuming that there are 10,000 training samples in total, to obtain 10 sub-training samples, 80% of the total training samples can be randomly s...

Embodiment 3

[0077] A model training device disclosed in this embodiment, such as Figure 4 As shown, the device includes:

[0078] A sampling module 410, configured to obtain at least one sub-training sample according to the training sample;

[0079] The single-model training and prediction module 420 is used to respectively train a plurality of machine learning models corresponding to each of the sub-training samples according to each of the sub-training samples, and obtain a corresponding machine learning model for each of the sub-training samples. predicted value;

[0080] A sample feature fusion module 430, configured to determine a fusion training sample according to the predicted value;

[0081] The target machine model training module 440 is configured to train the target machine learning model according to the fused training samples determined by the sample feature fusion module 430 .

[0082] optional, such as Figure 5 As shown, the target machine model training module 440 f...

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 application provides a model training method, and belongs to the field of computer technology. The method solves the problem that prediction results are inaccurate when models obtained by trainingof model training methods in the prior art are applied to application of data mining, searching or the like, and includes: obtaining at least one sub-training-sample according to a training sample; respectively training a plurality of machine learning models, which correspond to each sub-training-sample, according to each sub-training-sample, and acquiring prediction values of the corresponding machine learning models on each sub-training-sample; determining a fusion training sample according to the prediction values; and training a target machine learning model according to the fusion training sample. According to the model training method disclosed by the embodiment of the application, the target model is further trained through using prediction results of the models, which are obtainedby previous training, as a feature, and accuracy of a prediction effect of a model obtained by training can be effectively improved.

Description

technical field [0001] The present application relates to the field of computer technology, in particular to a model training method and device, and electronic equipment. Background technique [0002] As the amount of platform data increases, the utilization of platform data is particularly important. For example: modeling through platform data, using pre-trained models to predict user behavior or provide data that users are interested in. In the prior art, a common method is to pre-train a model and use the trained model to predict real-time data. Further, in order to improve the accuracy of the predicted data, there is a pre-training of multiple models in the prior art, and then each model is used to perform data prediction respectively, and finally, the prediction results are fused, for example, by combining each model The predicted scores are weighted and summed to obtain the final predicted score for the data. When training a model in the prior art, the preset dimens...

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): G06K9/62G06N3/08
CPCG06N3/08G06F18/214G06N20/20G06N20/10G06N5/01G06N7/01G06N3/045G06N20/00G06N5/04
Inventor 王子伟
Owner BEIJING SANKUAI ONLINE TECH CO LTD
Features
  • Generate Ideas
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com