Machine Learning Model Selection via Data Group Distance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Machine learning algorithms face challenges in improving prediction accuracy due to the lack of effective methods for selecting the most relevant data groups for predicting unknown data.
Innovation Solution
A data predicting method and apparatus that determines distances between predicting data and multiple data groups, selects a machine learning model corresponding to the data group with the shortest distance, and uses it for prediction, thereby improving accuracy, sensitivity, and specificity by clustering and training models on different data groups.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple machine learning models are trained on different data groups, then prediction accuracy is improved, but model selection complexity increases
Solution Approach 1:
The patent segments the training data into multiple distinct data groups, with each group used to train a separate machine learning model. This segmentation allows the system to specialize different models for different data characteristics, thereby improving prediction accuracy while managing complexity through organized model selection based on data group identification
Solution Approach 2:
The patent implements preliminary action by pre-calculating distances between data points and data groups before prediction is needed. These distance calculations are stored and reused during prediction, eliminating the need for repeated complex calculations and simplifying the model selection process at prediction time
2Measurement precision
If distance calculation between predicting data and multiple data groups is performed, then model selection accuracy is improved, but computational time increases
Solution Approach 1:
The patent performs distance calculations between data groups and potential prediction targets in advance, before actual prediction is needed. These pre-computed distances are stored for rapid retrieval during prediction, significantly reducing computational time while maintaining accurate model selection based on shortest distance criteria
Solution Approach 2:
The patent creates and stores copies of distance calculation results for reuse during prediction. Instead of recalculating distances between predicting data and all data groups each time, the system uses pre-computed distance information, reducing computational overhead while preserving the accuracy of model selection
Data Source
AI summary
A data predicting method and apparatus are provided. In the method, distances between a predicting data and multiple data groups are determined. A first machine learning model corresponding the data group having the shortest distance with the predicting data is selected from multiple machine learning models. The predicting data is predicted through the first machine learning model. Those machine learning models are trained by using different data groups, respectively.