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

Demand prediction method based on incremental algorithm

A demand forecasting and incremental technology, applied in computing, commerce, instruments, etc., can solve problems such as unquantifiable demand, achieve the effect of reducing impact and improving accuracy

Pending Publication Date: 2020-06-05
CENTRAL UNIVERSITY OF FINANCE AND ECONOMICS +1
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

None of the above algorithms mentioned factors that have an impact on demand but cannot be quantified or collected
These factors cannot be handled by feature engineering, and most of them cannot be solved by subdividing or summarizing requirements

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
  • Demand prediction method based on incremental algorithm
  • Demand prediction method based on incremental algorithm
  • Demand prediction method based on incremental algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] A large-scale 020 instant platform has a total of more than 100,000 stores, and the database has operation records of all stores from January 1, 2016 to August 31, 2019. Now it is planned to predict the order demand of 44 stores in the next day, and the forecast lead time τ=1 day. Through the preliminary analysis of historical sales, it is determined that the difference interval length δ = 1 day, and the training set time period does not exceed TP = [2018-11-1, 2017-11-2, ..., 2019-10-31] T . Some new stores will be launched after November 1, 2017, that is, the length of the time period is tp≤365.

[0052] Extract the prediction feature set S for store order demand prediction. Attribute features include historical sales, prices, promotions, reviews, page views, inventory, product attributes, store attributes, weather, holidays, and days of the week.

[0053] Extract the predictive features of store k for model training

[0054]

[0055] and demand vector wher...

Embodiment 2

[0068] A large supermarket has a total of 1,000 existing SKUs, and the database has operating records of all SKUs from January 1, 2016 to October 31, 2019. It is currently planned to conduct demand forecasting for SKUs with a procurement lead time of 3 days to support daily replenishment decisions. Forecast lead time τ = 3 days. Through the preliminary analysis of historical sales, it is determined that the difference interval length δ = 7 days, and the training set time period is TP = [2016-1-1, 2017-1-2, ..., 2019-10-31] T , the time period length is tp, a total of 1399 days.

[0069] Extract the prediction feature set S for product demand prediction. Attribute features include historical sales, prices, promotions, reviews, page views, inventory, product attributes, competing product information, page display, weather, holidays, and days of the week.

[0070] Extract the predictive features of product k for model training

[0071]

[0072] and demand vector where ...

Embodiment 3

[0085] A large supermarket has a total of 1,000 existing SKUs, and the database has operating records of all SKUs from January 1, 2016 to October 31, 2019. Now it is planned to forecast the demand for the 30-day sales in October 2019 of SKUs with a procurement lead time of 7 days in a region to support the replenishment decision of the general warehouse. Forecast lead time τ = 7 days. Through the preliminary analysis of historical sales, it is determined that the difference interval length δ = 1 month, and the training set time period is TP = [2016-1-1, 2017-1-2, ..., 2019-10-31] T , the time period length is tp, a total of 1399 days.

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 demand prediction method based on an incremental algorithm. The demand prediction method comprises the following steps of: extracting features and then carrying out differential processing on the features to obtain increments of the current features relative to historical features; carrying out differential processing on the demand to obtain an incremental demand of the current demand relative to a historical demand; constructing a feature matrix by means of the incremental features and the initial features; forming training data by using the feature matrix and the increment demand; training an incremental demand estimation model by adopting a machine learning method; and estimating the future increment by using the learned model so as to obtain future demand prediction. According to the demand prediction method, the influence of factors which cannot be quantitatively characterized on prediction can be reduced by predicting the demand increment, and the accuracy of demand prediction is effectively improved.

Description

technical field [0001] The invention belongs to the interdisciplinary field of machine learning and supply chain management, and in particular relates to a method for pattern mining and real-time prediction of demand for products or services, specifically a demand forecasting method based on an incremental algorithm. Background technique [0002] The development of the Internet, mobile marketing, and new retail has put forward higher requirements for the sensitivity of consumer demand perception, inventory response speed, and supply chain management efficiency. In the field of supply chain management, inventory management and decision-making occupy a very important position; and accurately predicting future demand and demand changes will greatly improve the operating level of inventory. How to effectively predict the future demand of new products, especially the real-time demand, has gradually become an important topic and difficult problem in product operation and supply ch...

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): G06Q30/02
CPCG06Q30/0202
Inventor 代宏砚周伟华肖沁周云
Owner CENTRAL UNIVERSITY OF FINANCE AND 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