Kernel Methods TRIZ Cases
- Machine Learning Model Classification Accuracy via Reliability Voting
A machine learning model generates classification output values and feature spectra to determine class labels through reliability-based voting.
- Kernel-Based Data Screening for Industrial Machine Model Building
Unsupervised kernel methods automate reference data selection, resolving labor-intensive manual screening bottlenecks.
- Variable Importance Ranking in Nonparametric Machine Learning Models
A statistical system ranks variables by binning raw values into equally sized buckets and assigning rank numbers based on mean probability values.