Explainable Decision Trees for User Similarity in Software Recommendations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing software applications face inefficiencies in delivering targeted messaging due to difficulties in identifying similar users, leading to irrelevant recommendations and wastage of resources, as current machine learning models lack explainability and accuracy in user classification.
Innovation Solution
The method involves generating decision trees from transaction history data to represent users, allowing for the identification of similar users based on transaction patterns with counterparties, and using these trees to calculate distance metrics for personalized recommendations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If random message selection is used, then implementation simplicity is maintained, but message relevance to users deteriorates
Solution Approach 1:
The patent replaces the mechanical/random message selection system with an information-processing system using decision trees and similarity metrics. Instead of randomly selecting messages, the system calculates similarity between users based on transaction history and counterparty relationships, then delivers targeted messages. This substitution transforms a simple but ineffective mechanical process into a complex but effective information-based process.
2Extent of automation
If traditional machine learning models are used for user classification, then automation is achieved, but explainability deteriorates
Solution Approach 1:
The patent creates explicit copies of user characteristics through decision trees that replicate the reasoning process. Instead of using opaque machine learning models, the system builds decision trees that copy and represent the classification logic in an interpretable format. These trees can be visualized and explained, showing exactly how users are classified into similar groups based on their transaction patterns.
3Reliability
If targeted messaging based on user similarity is implemented, then message relevance improves, but computing resource consumption increases
Solution Approach 1:
The patent segments the user base into distinct groups using decision trees based on transaction history and counterparty relationships. By dividing users into similar-groups rather than treating all users individually, the system reduces the computational burden of message delivery. Once users are segmented, messages can be efficiently targeted to groups rather than requiring individualized processing for each user.
4Measurement precision
If comprehensive user analysis is performed to identify similar users, then recommendation accuracy improves, but processing time increases
Solution Approach 1:
The patent performs preliminary actions by pre-building decision trees and calculating user similarities in advance. The system processes transaction history data beforehand to create user profiles and identify similar users before message delivery is needed. This preliminary processing allows for faster real-time message delivery while maintaining high accuracy in user matching.
Data Source
AI summary
Certain aspects of the present disclosure provide techniques for training decision trees representing users of a software application. An example method generally includes generating, from a transaction history data set for a plurality of users of a software application, a plurality of grouped data sets including transactions grouped by counterparty. A plurality of feature vectors are generated from the plurality of grouped data sets. Each feature vector generally corresponds to a user of the plurality of users and includes a plurality of features describing relationships between the user and a plurality of counterparties in a transaction history associated with the user. A decision tree is trained based on the plurality of feature vectors. The decision tree generally includes a plurality of paths terminating in a similar or different classification, and the plurality of paths distinguishes a user associated with the decision tree from other users of the software application.