Method and system for generating ensemble demand forecasts

Patent metadata

Field	Value
Publication number	US-11373199-B2
Application number	US-201816172575-A
Country	US
Kind code	B2
Filing date	Oct 26, 2018
Priority date	Oct 26, 2018
Publication date	Jun 28, 2022
Grant date	Jun 28, 2022

Abstract

Methods and systems for forecasting demand for a plurality of items are provided. In particular, the demand forecasting system and methods described herein are useful for predicting demand of products in a retail context. Forecast models are built and used to score incoming sales data to predict future demand for items. Forecast models are validated by evaluating actual demand against predicted demand and using that information to inform how future ensemble forecast will be generated. Forecasts may be broken down into smaller components to satisfy a variety of requests for data from client applications.

First claim

The invention claimed is:

1. A system for forecasting demand for a plurality of items, the system comprising:
a common data preparation engine comprising:
a computing system including a standard data store, processor, and memory communicatively coupled to the processor, the memory storing instructions executable by the processor to:
continually receive sales data for a plurality of items sold within a retail enterprise;
incrementally process the sales data as it is received to convert the sales data to common format data; and
store the common format data in the standard data store;

an enterprise forecast engine comprising:
a computing system including a processor, and a memory communicatively coupled to the processor, the memory storing instructions executable by the processor to:
receive parameters for a demand forecast defining at least one of: one or more items of the plurality of items, a time period, and one or more locations of the retail enterprise;
access past forecasting performance data of a plurality of component models for the received parameters;
build an ensemble forecasting model using at least two of the plurality of component models, the at least two of the plurality of component models selected and weighted based on the past forecasting performance data;
access the common format data;
generate an aggregate demand forecast using the ensemble forecasting model and the common format data;
communicate the aggregate demand forecast to a forecasts data store for storage; and
continually update the ensemble forecasting model in response to real-time updates to the common format data and time period-based demand changes by modifying selection and weighting of the at least two of the plurality of component models used to build the ensemble forecasting model; and

a server comprising an API accessible by one or more client applications, the API configured to:
receive requests for demand forecasts,
communicate demand forecast requests to the forecasts data store, and
receive demand forecasts responsive to the demand forecast requests from the forecasts data store.

2. The system of claim 1 , wherein the ensemble forecasting model is built from the at least two of the plurality of component models selected and weighted using an affine function.

3. The system of claim 1 , wherein the at least two of the plurality of component models include one or more of an autoregressive integrated moving average (ARIMA) model or a locally estimated scatterplot smoothing (LOESS) model.

4. The system of claim 1 , wherein the API is further configured to generate an administrator user interface for visualizing demand forecast data.

5. The system of claim 1 , further comprising a cloud platform with at least one load balancer to direct demand forecast requests to one of a plurality of servers, including the server, for processing, the direction based on which of the plurality of servers is least busy processing other demand forecast requests.

6. The system of claim 1 , wherein the past forecasting performance data is evaluated with a forecast validation engine by running multiple forecasts for one set of sales data and cross-validating distributions of results.

7. The system of claim 1 , wherein the common data preparation engine is further configured to continually receive data, including the sales data and additional data, having different formats from a plurality of different sources and reformat the received data into the common format data to enable use of the data by the ensemble forecasting model.

8. The system of claim 7 , wherein the additional data comprises at least one of the following: catalog data, location data, inventory data, promotion data, and planogram data.

9. The system of claim 1 , wherein the past forecasting performance takes into account seasonal changes in demand.

10. The system of claim 1 , further comprising a cumulative distribution function (CDF) service configured to convert the aggregate demand forecast into forecast distributions.

11. The system of claim 1 , wherein the time period-based demand changes include at least one of demand changes based on seasonality and demand changes based on item promotions.

12. A method of forecasting demand for a plurality of items within a retail enterprise, the method comprising:
continually receiving sales data at a common data preparation engine, the sales data comprising past sales data and current sales data for sales of the plurality of items at locations of the retail enterprise, including physical stores and online;
incrementally processing the sales data as it is received to convert the sales data to common format data;
receiving parameters for a demand forecast, the parameters defining at least one of: one or more items of the plurality of items, a time period, and one or more locations of the retail enterprise;
building, with an enterprise forecast engine, an ensemble forecast model by combining two or more of a plurality of component models, the two or more of the plurality of component models selected and weighted based on an assessment of past forecasting performance of the plurality of component models for the received parameters;
continually updating the ensemble forecast model in response to real-time updates to the common format data and time period-based demand changes by modifying a selection and weighting of the two or more of the plurality of component models used to build the ensemble forecast model;
analyzing, with the ensemble forecast model, the common format data to generate an aggregate demand forecast; and
storing the aggregate demand forecast in a forecasts data store.

13. The method of claim 12 , further comprising:
receiving, at an API, a request from a client application for a demand forecast for one or more items of the plurality of items;
querying the forecasts data store for an aggregate demand forecast corresponding to the request; and
responding to the client application with the requested demand forecast.

14. The method of claim 13 , wherein the requested demand forecast is for an individual item, over a 1 week time period, for all sales of the retail enterprise.

15. The method of claim 13 , wherein the request further comprises a time period, and a location of the retail enterprise.

16. The method of claim 13 , wherein the request is communicated to a resource manager associated with a query service, and the query service accesses the aggregate demand forecast from the forecasts data store that matches the request.

17. The method of claim 13 , wherein the requested demand forecast is visualized on a user interface for viewing and analysis by an administrator user.

18. The method of claim 12 , further comprising, as part of the selection and weighting of each of the two or more of the plurality of component models used to build the ensemble forecast model, validating each of the two or more of the plurality of component models.

19. A non-transitory computer-readable storage medium comprising computer-executable instruction which, when executed by a computing system, cause the computing system to perform a method of generating a demand forecast for one or more items, the method comprising:
continually receiving store sales data and web sales data for a plurality of items sold within a retail enterprise at a common data preparation engine in real time;
incrementally processing the store sales data and web sales data to convert the store sales data and web sales data to common format data;
receiving parameters

Assignees

• Target Brands Inc

Inventors

• Morgan Adam
• Das Anand
• Shah Samir H
• Yelland Phillip

Classifications

• G06N3/045

  Combinations of networks · CPC title

• G06N3/044

  Recurrent networks, e.g. Hopfield networks · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/0442

  characterised by memory or gating, e.g. long short-term memory [LSTM] or gated recurrent units [GRU] · CPC title

• G06Q30/0202 Primary

  Market predictions or forecasting for commercial activities · CPC title