Counting Machine for Personalized Entity Ratings via Minkowski Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for decision-making involving multiple criteria and ordered categorical values, such as MCDM/MCDA, face challenges in providing consumers with a realistic, meaningful, and actionable guide due to the scarcity of domination relationships among entities and the impracticality of pairwise comparisons, especially when dealing with a large number of entities, leading to suboptimal solutions.
Innovation Solution
A counting machine employing Minkowski-distance semi-supervised machine learning to generate personalized ratings that are secure and conceal the underlying entity ratings, allowing users to rank their top and bottom preferences with adaptive menus and flexible selection options, preventing back-solving and reverse engineering.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If MCDM/MCDA methods are used to evaluate entities with multiple criteria, then comprehensive analysis is achieved, but the results are not actionable and do not provide clear guidance to consumers
Solution Approach 1:
The patent transforms the output format of MCDM/MCDA results from a set of ratings into a single actionable recommendation. This is achieved by changing the parameter representation from multiple dimensional ratings to a unified recommendation indicator that directly guides consumer decision-making, resolving the contradiction between comprehensive evaluation and actionable output.
2Measurement precision
If pairwise comparisons are performed among all entities, then complete ranking is achieved, but the computational complexity becomes impractical for large numbers of entities
Solution Approach 1:
The patent applies preliminary filtering based on user preferences before performing pairwise comparisons. By pre-identifying relevant entities according to user-specified criteria and preferences, the system reduces the number of entities requiring full pairwise comparison, thereby maintaining ranking completeness while significantly reducing computational complexity.
3Loss of information
If all factor ratings are disclosed to consumers, then transparency is achieved, but the commercial value of the ratings is compromised
Solution Approach 1:
The patent extracts only the essential information needed for consumer decision-making (the final recommendation) while leaving the detailed factor ratings confidential. This selective extraction provides sufficient transparency for users to make informed choices while preserving the commercial value of the underlying rating data by not disclosing all individual factor scores.
4Quantity of substance
If multiple factor ratings are provided to consumers, then detailed information is available, but consumers cannot process and act on the information effectively
Solution Approach 1:
The patent extracts the most critical decision-making information from the set of multiple factor ratings and presents it in a simplified, actionable format. By taking out only the essential recommendation rather than presenting all factor ratings, the system maintains information quality while dramatically improving ease of processing and actionability for consumers.
Data Source
AI summary
A machine for choosing among possible entities is disclosed comprising an electronic device with a display, a user interface and a memory storage storing a table listing predetermined attribute preferences associated with index numbers computed from user rankings. Influencing factors are displayed and partially ranked by the user. The top influencing factors are selected from a first menu and the least important factors chosen in a second menu which is adaptive to the first menu selections. The machine further includes a counting circuit, and an input-output device driver. The input-output device driver transmits the top ranks and the bottom ranks to the counting circuit, which determines the index number corresponding to the mapping of the top ranks and the bottom ranks. The memory storage unit retrieves the predetermined entity ranking associated with the mapping, derived from experts in the field, opinions of consumers, and machine learning, and displays it.