Prediction Apparatus Tree Structure Rule Aggregation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing prediction technologies face inefficiencies in predicting optimal actions due to the need to try all classification rules, leading to increased processing costs.
Innovation Solution
A prediction program and apparatus that generates tree structure data based on rule information and determines the degree of contribution of attribute values to a predetermined label, allowing for efficient prediction by aggregating hypotheses and applying a predetermined order condition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all classification rules are tried to predict optimal actions, then prediction accuracy is improved, but processing costs increase
Solution Approach 1:
The patent segments the classification rules into a hierarchical tree structure with multiple levels. Instead of evaluating all rules equally, the system divides them into parent nodes and child nodes, allowing selective evaluation starting from parent nodes. This segmentation enables the system to achieve accurate predictions by traversing only necessary paths in the tree, thereby reducing processing costs while maintaining prediction accuracy.
Solution Approach 2:
The patent performs preliminary action by pre-processing the classification rules into a structured tree format before actual prediction. The tree structure is built in advance with parent-child relationships established, so that during prediction, the system can efficiently navigate the pre-organized structure without needing to evaluate all rules from scratch. This preliminary structuring reduces the computational burden during the actual prediction phase.
2Measurement precision
If all classification rules are tried to predict optimal actions, then prediction accuracy is improved, but processing time increases
Solution Approach 1:
The patent segments the classification rules into a hierarchical tree structure with multiple levels. Instead of evaluating all rules equally, the system divides them into parent nodes and child nodes, allowing selective evaluation starting from parent nodes. This segmentation enables the system to achieve accurate predictions by traversing only necessary paths in the tree, thereby reducing processing time while maintaining prediction accuracy.
Solution Approach 2:
The patent performs preliminary action by pre-processing the classification rules into a structured tree format before actual prediction. The tree structure is built in advance with parent-child relationships established, so that during prediction, the system can efficiently navigate the pre-organized structure without needing to evaluate all rules from scratch. This preliminary structuring reduces the computational burden during the actual prediction phase.
Data Source
AI summary
A non-transitory computer-readable recording medium stores therein a prediction program that causes a computer to execute a process including receiving input data to be predicted, generating a tree structure data based on a plurality of pieces of rule information each indicated by an association of a combination of attribute values of a plurality of attributes with a label according to a predetermined order condition for the plurality of attributes, the tree structure data being obtained by aggregating the plurality of pieces of rule information, the tree structure data including an attribute value as a branch, and determining a degree of contribution to make a determination result on a predetermined value of a predetermined attribute reach a predetermined label based on a likelihood of obtaining a value of the predetermined label as the determination result, when the attribute value of the predetermined attribute is determined to be the predetermined value.