Demand Prediction Model Updating via Abnormal Error Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing demand prediction methods rely heavily on user expertise to select explanatory variables, leading to potential omission of necessary variables and reduced prediction accuracy, as they do not automatically account for abnormal values or update models efficiently.
Innovation Solution
A demand prediction method that calculates errors between predicted and actual past demand values, determines abnormal values, acquires new explanatory variables, and updates the prediction model using these variables to improve future demand forecasting, thereby reducing user input operations and increasing accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If user expertise is relied upon to select explanatory variables, then the model can be constructed with available knowledge, but the prediction accuracy deteriorates due to potential omission of necessary variables
Solution Approach 1:
The system automatically detects abnormal values in prediction errors and autonomously acquires new explanatory variables without requiring continuous user intervention. The error detection unit, variable acquisition unit, and model update unit work together to enable the system to self-improve its prediction model by identifying missing variables through abnormal error analysis and automatically incorporating them into the prediction model.
Solution Approach 2:
The system implements a feedback mechanism where prediction errors are continuously monitored, abnormal values are detected, and this information feeds back into the variable acquisition process. The error detection unit provides feedback about prediction accuracy, which triggers the acquisition of new explanatory variables, which then updates the model to improve future predictions, creating a closed-loop system for continuous improvement.
2Productivity
If manual acquisition of explanatory variables is performed, then variables can be selected based on user judgment, but the time required to acquire variables increases
Solution Approach 1:
The system performs preliminary analysis of prediction errors to detect abnormal values before formal model updates are required. By continuously monitoring errors and pre-identifying when new explanatory variables are needed, the system prepares for model updates in advance, reducing the actual time required for variable acquisition and model reconstruction when improvements are needed.
Solution Approach 2:
The automated error detection and variable acquisition system eliminates the need for manual user intervention in identifying and selecting new explanatory variables. The system serves itself by automatically detecting when prediction accuracy deteriorates due to missing variables and autonomously acquiring appropriate variables from available data sources, significantly reducing the time investment required.
3Measurement precision
If the prediction model is updated frequently with new variables, then prediction accuracy improves, but the number of user input operations increases
Solution Approach 1:
The system uses feedback from prediction error analysis to determine when model updates are necessary. Rather than updating the model on a fixed schedule or in response to every minor error fluctuation, the feedback mechanism monitors error patterns and triggers updates only when abnormal values indicate that new explanatory variables are needed, minimizing unnecessary user input operations while maintaining prediction accuracy.
Solution Approach 2:
The system performs self-updates by automatically acquiring and integrating new explanatory variables without requiring user input for each update operation. The variable acquisition unit autonomously identifies and incorporates new variables, and the model update unit automatically retrain's the model, eliminating the need for users to manually initiate or manage each model update process.
Data Source
AI summary
A demand prediction method using a computer includes calculating an error between a prediction value of an amount of a past demand and a measurement value of the amount of the past demand, the prediction value of the amount of the past demand being calculated by inputting a measurement value of a past explanatory variable to a prediction model that is constructed in accordance with a measurement value of an amount of a demand at a predetermined location and an explanatory variable that serves as an external factor that affects an increase or a decrease in the amount of the demand at the predetermined location, determining whether the calculated error is an abnormal value, acquiring a new explanatory variable if the error is determined to be the abnormal value, updating the prediction model in accordance with the acquired new explanatory variable, and newly predicting an amount of a future demand using the updated prediction model.