Intransitive Choice Prediction via Relativized Matrix
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional discrete choice models struggle to predict intransitive choices made by individuals, particularly when dealing with large quantities of data involving many consumers and products, as they fail to stably estimate model parameters effectively.
Innovation Solution
A calculation apparatus and method that acquire feature vectors for alternatives, calculate absolute and relative evaluations using a relativized matrix, and adjust parameters to improve relative evaluation accuracy, enabling the prediction of intransitive choices by considering attraction and compromise effects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional discrete choice models are used to predict choices, then the model structure is simple and easy to implement, but the model cannot stably estimate parameters for large quantities of data and fails to predict intransitive choices
Solution Approach 1:
The patent segments the evaluation process into two distinct components: absolute evaluation (independent of alternative combinations) and relative evaluation (dependent on alternative combinations). This segmentation allows the model to separately handle the complexity of intransitive choices while maintaining computational feasibility through structured decomposition of the evaluation function.
Solution Approach 2:
The patent introduces a new dimension by adding the relativized matrix that represents the influence of alternative combinations on evaluations. This transforms the traditional single-dimension utility function into a multi-dimensional framework that captures both absolute preferences and context-dependent relative preferences, enabling stable parameter estimation for intransitive choices.
2Adaptability or versatility
If conventional discrete choice models are used, then the calculation process is simple, but the model cannot account for attraction effects and compromise effects in actual choices
Solution Approach 1:
The patent performs preliminary calculation of absolute evaluation vectors that are independent of alternative combinations. By pre-computing these absolute evaluations, the model efficiently handles the complexity of accounting for attraction and compromise effects without requiring complex iterative calculations for each choice scenario.
Solution Approach 2:
The relativized matrix serves as an intermediary that bridges absolute evaluations and final relative evaluations. This intermediary component captures the influence of alternative combinations (including attraction and compromise effects) and transforms absolute evaluations into context-dependent relative evaluations, enabling the model to account for complex behavioral effects.
3Measurement precision
If large quantities of learning data are used, then the model can capture more consumer behavior patterns, but conventional models cannot stably estimate parameters
Solution Approach 1:
The patent segments the parameter estimation process into separate estimation of absolute evaluation parameters and relativized matrix parameters. This segmentation allows each component to be estimated more reliably from large datasets independently, improving overall parameter estimation stability while capturing detailed consumer behavior patterns.
Solution Approach 2:
The patent adds the relativized matrix dimension that explicitly models the influence of alternative combinations. This additional dimension allows the model to capture nuanced consumer behavior patterns from large datasets while maintaining parameter estimation stability through the structured mathematical framework that separates absolute and relative evaluation components.
Data Source
AI summary
A calculation apparatus includes a feature vector acquisition unit for acquiring a feature vector that corresponds to each alternative of a plurality of choice sets; an absolute evaluation calculation unit for calculating an absolute evaluation vector that represents an absolute evaluation of alternatives independent of a combination of the plurality of alternatives; a relativized-matrix calculation unit for calculating a relativized matrix that represents relative evaluations of the plurality of alternatives in a choice set; and a relative evaluation calculation unit for calculating a relative evaluation vector that represents a relative evaluation of each alternative of the plurality of alternatives from a product of multiplying the relativized matrix by the absolute evaluation vector. The calculation apparatus can predict intransitive choices of a person who exhibits intransitive preferences.