ML Feature Clustering for Recommendation Accuracy
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing machine learning-based item recommendation systems often have low success rates due to generic recommendations lacking customization, leading to inefficient use of computing resources and low acceptance rates.
Innovation Solution
Implementing resource-efficient model feature evaluation and clustering techniques to generate customized recommendations by identifying key features and user attributes, reducing the feature set using Shapley values and correlation analysis, and deploying a predictive model with a clustering algorithm like k-means to tailor offers based on user interactions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If generic recommendations are generated without customization, then computing resources are consumed significantly, but acceptance rates remain low
Solution Approach 1:
The patent segments users into distinct clusters based on their attributes and behaviors using clustering algorithms. This segmentation allows the system to generate customized recommendations for each cluster rather than generic recommendations for all users, thereby improving acceptance rates while optimizing resource usage by targeting specific segments.
Solution Approach 2:
The patent applies local quality by tailoring recommendation characteristics to specific user clusters. Each cluster receives recommendations with attributes customized to that cluster's preferences and behaviors, rather than applying a uniform recommendation approach to all users. This localized customization improves acceptance rates while avoiding the waste of generating inappropriate generic recommendations.
2Productivity
If feature sets are reduced using Shapley values and correlation analysis, then computational efficiency improves, but model accuracy may be compromised
Solution Approach 1:
The patent extracts and removes redundant or low-contributing features from the feature set using Shapley values and correlation analysis. By identifying and eliminating features that have minimal impact on model predictions or are highly correlated with other features, the system reduces computational complexity while preserving the essential information needed for accurate predictions.
Solution Approach 2:
The patent changes the parameter set by selecting a optimized subset of features based on their contribution to model output. This parameter selection process, guided by Shapley values and correlation analysis, transforms the original large feature set into a more efficient subset that maintains predictive accuracy while reducing computational burden.
3Productivity
If customized recommendations are generated for each user, then acceptance rates improve, but computing resources are consumed excessively
Solution Approach 1:
The patent segments the user base into clusters with similar attributes and behaviors. Instead of generating fully customized recommendations for each individual user, the system generates customized recommendations for each cluster. This approach maintains high acceptance rates by providing relevant recommendations while significantly reducing system complexity compared to individual-level customization.
Solution Approach 2:
The patent applies universality by creating cluster-level customized recommendations that serve multiple users within each cluster. Each cluster receives a customized recommendation strategy that works effectively for all members of that cluster, providing a universal solution for each segment while avoiding the excessive complexity of individual customization.
Data Source
AI summary
The present disclosure generally relates to systems, software, and computer-implemented methods for using resource-efficient model feature evaluation and clustering techniques to refine outputs of machine learning models. One example method includes receiving a set of data relating to a user and a particular item. The set of data can be input to a predictive model. A model output specifying a particular likelihood that the user will obtain the particular item can be obtained from the predictive model. Scores for a set of features of the predictive model can be computed based on the model output. A cluster can be identified from among a plurality of clusters using a clustering model. The customized recommendation can be generated for the user to obtain the particular item based on the identified cluster. The customized recommendation can be transmitted via a network interface and to a device corresponding to the user.