Tree-Based Power Demand Estimation Using Goodness-of-Fit Branching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing tree structure generation methods, such as CART and CHAID, face challenges in accurately estimating future power demand due to limitations in branching conditions, leading to difficulties in generating deeper tree structures and thus less accurate predictions.
Innovation Solution
A system generates a first tree structure representing measurement data sets and calculates goodness-of-fit data for attribute items, allowing for the association of branch conditions with branch portions, enabling the creation of a second tree structure for improved data estimation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If general tree structure generation methods (CART, CHAID) are used, then the tree structure can be generated quickly, but the tree structure cannot go deeper when branch conditions are not found, resulting in lower estimation accuracy
Solution Approach 1:
The patent pre-generates a tree structure based on measurement data sets before associating branch conditions. This preliminary tree structure provides a framework that can later be enhanced with branch conditions, allowing the system to start with a basic structure and progressively improve it without being constrained by the need to find perfect branch conditions upfront.
Solution Approach 2:
The patent calculates goodness-of-fit data to evaluate how well branch conditions match the tree structure, and uses this feedback to iteratively improve the association between branch conditions and branch portions. This feedback mechanism allows the system to continuously refine the tree structure and branch conditions to achieve better estimation accuracy.
2Measurement precision
If deeper tree structures are generated to improve estimation accuracy, then prediction precision improves, but the complexity of finding appropriate branch conditions increases
Solution Approach 1:
The patent generates the tree structure based on measurement data sets before attempting to associate branch conditions. This preliminary structure establishes the hierarchical framework first, making it easier to subsequently identify and associate relevant branch conditions at each level without the overwhelming complexity of doing both simultaneously.
Solution Approach 2:
The patent introduces goodness-of-fit data as an intermediary metric to evaluate and guide the association between branch conditions and branch portions. This intermediary measure simplifies the complex task of branch condition identification by providing a quantitative criterion for assessing matches, making the process more systematic and less difficult.
3Device complexity
If branch conditions are strictly required for each branch portion, then the tree structure remains simple, but nodes cannot be decided when branch conditions are not found, limiting tree depth
Solution Approach 1:
The patent makes the tree structure dynamic by allowing branch conditions to be optionally associated with branch portions based on goodness-of-fit evaluation. Rather than requiring rigid branch conditions at every node, the system adaptively determines where branch conditions are appropriate, enabling the tree to expand to appropriate depths while maintaining simplicity where conditions are not available.
Solution Approach 2:
The patent changes the parameter of branch condition presence from a binary requirement to a probabilistic association based on goodness-of-fit scores. This parameter change allows the system to flexibly adjust the depth and complexity of the tree structure based on the quality of available branch conditions, balancing simplicity and accuracy.
Data Source
AI summary
A system generates a first tree structure representing a relation of a plurality of measurement data sets, and generates goodness-of-fit data based on at least a part of attribute data regarding one or more branch portions included in the first tree structure. The attribute data includes one or more attribute values at one or more points in time regarding each of one or more attribute items. The goodness-of-fit data includes goodness-of-fit for each attribute item regarding each branch portion. With regard to each attribute item for each branch portion, goodness-of-fit is a value calculated based on a parent node and two or more child nodes belonging to the relevant branch portion, and one or more attribute values corresponding to the relevant attribute item, and represents a degree that the relevant attribute item will fit as a base of a branch condition. The system generates a second tree structure in which a branch condition decided based on the goodness-of-fit data is associated with a branch portion included in the first tree structure, and performs data estimation using the second tree structure.