Learning Model Construction via Hierarchical Algorithm Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Constructing multiple learning models with different structures to optimize label prediction accuracy is time-consuming, as existing methods require executing all possible algorithms, leading to inefficiency and unnecessary computations.
Innovation Solution
An information processing apparatus that acquires learning data, determines hierarchical relationships among objective variables, and selects appropriate construction algorithms based on these relationships to construct a learning model efficiently, reducing the need to execute all algorithms and allowing for accurate prediction of construction time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple learning models with different structures are constructed to optimize label prediction accuracy, then prediction accuracy is improved, but construction time increases
Solution Approach 1:
The patent changes the parameter of algorithm selection based on the hierarchical relationship of objective variables. By analyzing the hierarchy structure, the system dynamically selects appropriate construction algorithms rather than executing all possible algorithms, thus achieving accurate predictions while reducing construction time through parameter-based optimization
Solution Approach 2:
The patent segments the algorithm execution process by dividing algorithms into those that must be executed and those that can be skipped based on hierarchical relationships. This segmentation allows the system to construct multiple learning models with different structures (for improved accuracy) while avoiding unnecessary computations (reducing time loss)
2Reliability
If all construction algorithms are executed to ensure comprehensive model construction, then model completeness is improved, but computational efficiency deteriorates
Solution Approach 1:
The patent performs preliminary analysis of the hierarchical relationship among objective variables before executing construction algorithms. This preliminary action enables the system to determine which algorithms are necessary and which can be skipped, ensuring model completeness is maintained while computational efficiency is improved by avoiding unnecessary algorithm executions
Solution Approach 2:
The patent extracts and removes unnecessary construction algorithms from the execution list based on hierarchical relationship analysis. By taking out algorithms that are not needed for a given hierarchy structure, the system maintains model completeness for relevant patterns while significantly improving computational efficiency by eliminating redundant calculations
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Learning data representing the relationship between explanatory variables and objective variables is acquired by an acquirer. In the learning data acquired by the acquirer, a hierarchical relationship among a plurality of items included in the objective variable is determined by a hierarchy determiner. A construction algorithm to be executed among a plurality of construction algorithms for construction of a learning model is determined by an algorithm determiner based on the hierarchical relationship determined by the hierarchy determiner. A first learning model is constructed by execution of the construction algorithm determined by the algorithm determiner by a learner.