Decision Tree Generation Using User Answer Reliability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing decision tree generation methods fail to account for the possibility of user errors in answers, leading to suboptimal inquiry order determination and increased number of inquiries in dialog systems.
Innovation Solution
A decision tree generating apparatus that calculates information gain based on entropy reduction and user answer reliability, recursively determining attribute nodes to minimize inquiries and errors, even when user answers are incorrect.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If decision tree is generated without considering user answer reliability, then the generation process is simple, but the number of inquiries increases and classification accuracy decreases
Solution Approach 1:
The patent modifies the information gain calculation by introducing reliability as a new parameter. The information gain is recalculated as: information gain = entropy reduction × reliability. This parameter change allows the decision tree to prioritize attributes that are both informative and reliable, reducing the number of inquiries needed while maintaining or improving classification accuracy.
Solution Approach 2:
The patent implements feedback by using the calculated reliability values to adjust the inquiry strategy. The system continuously evaluates user answer reliability and uses this feedback to optimize the decision tree structure, selecting attributes that minimize expected inquiries while accounting for potential user errors.
2Ease of manufacture
If traditional information gain based on entropy reduction alone is used, then the calculation is straightforward, but the inquiry order is suboptimal when user errors occur
Solution Approach 1:
The patent extends the traditional information gain formula by multiplying entropy reduction by reliability: information gain = entropy reduction × reliability. This modification maintains computational tractability while significantly improving inquiry efficiency by accounting for user answer reliability in the attribute selection process.
3Device complexity
If decision tree does not account for user answer errors, then the structure is simpler, but more inquiries are needed to achieve accurate classification
Solution Approach 1:
The patent incorporates reliability as a weighting parameter in the information gain calculation, modifying the decision tree construction criterion from pure entropy reduction to reliability-weighted information gain. This allows the tree structure to adapt to user answer reliability without fundamentally changing the decision tree algorithm, maintaining simplicity while improving efficiency.
Data Source
AI summary
A decision tree generating apparatus includes an information gain calculator and a decision tree generator. When a classification target data set including a plurality of pieces of classification target data respectively having different attributes with attribute values assigned thereto is segmented into subsets in a form of a decision tree, the information gain calculator calculates an amount of entropy reduction on each attribute, and calculates an information gain, based on the amount of reduction in the entropy and reliability of a user's answer responsive to an inquiry asking about the attribute. The decision tree generator successively determines an attribute having a maximum information gain to be a node of the decision tree by recursively iterating the segmentation of the pre-segmentation data set, and generates the decision tree that is to be used to determine an order of the inquiries.