Demand Prediction Using Trend Pattern Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques face difficulties in performing demand prediction with variations, such as optimal, upside, or downside trends, as they do not support predictions effectively in terms of demand changes over different time periods.
Innovation Solution
A demand prediction apparatus that acquires changes in demand over various time periods, identifies trend pattern parameters, and evaluates suitability to select appropriate trend patterns for predicting demand variations, including optimal, upside, and downside predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional demand prediction techniques are used, then the prediction model is simple, but the prediction cannot capture demand variations (optimal, upside, downside trends)
Solution Approach 1:
The patent segments the demand prediction process into multiple independent components: trend pattern generation unit, parameter identification unit, and selection unit. Each component handles a specific aspect of prediction, allowing the system to capture different demand variations (optimal, upside, downside) through separate trend patterns while maintaining overall system manageability through modular architecture.
Solution Approach 2:
The system dynamically adjusts the prediction approach by selecting different trend patterns based on identified parameters and evaluated suitability. The parameter identification unit determines optimal parameters from multiple combinations, and the selection unit chooses appropriate trend patterns based on demand variation characteristics, enabling the system to adapt to different prediction scenarios without requiring a completely different model for each case.
2Measurement precision
If multiple trend patterns and time periods are analyzed, then prediction accuracy for demand variations improves, but computational complexity increases
Solution Approach 1:
The trend pattern generation unit pre-generates multiple trend patterns and their corresponding parameters before the actual prediction process. This preliminary preparation allows the parameter identification unit to efficiently select from pre-computed patterns rather than calculating everything in real-time, reducing computational complexity while maintaining high prediction accuracy through comprehensive pattern analysis.
Solution Approach 2:
The selection unit uses evaluation results as feedback to determine which trend patterns and parameters to select for final prediction. By evaluating suitability based on demand variations and using this feedback to guide parameter selection, the system achieves high prediction accuracy while avoiding unnecessary computational effort by focusing only on the most relevant patterns identified through the evaluation process.
Data Source
AI summary
A demand prediction apparatus allows for a prediction with a variation in terms of demand by acquiring, out of changes in demand in a first time period, changes in the demand in a second time period which is a part of the first time period acquiring changes in the demand in third time periods, which is a plurality of time periods excluding the second time period within the first time period, for each of the third time periods having different lengths, identifying, for combinations of a plurality of trend patterns and the third time periods, parameters of the corresponding plurality of trend patterns such that a value of a trend pattern of the plurality of trend patterns is equal to the demand at a certain time point in the first time period, evaluating suitability with the changes in the demand in the second time period for each of the plurality of trend patterns for each of the combinations in which a parameter of the parameters is identified, and selecting one or more trend patterns from among the plurality of trend patterns for each of the combinations based on evaluation results.