DIStIL
Domain Informed Structural Invariant Learner
- DIStIL is an automated data sifting and feature extraction tool
- Designed explicitly for high-dimensional datasets
- Identify the most contrasting samples that contribute to the core features
- Automatically rectifies number of over-specified extracted features
- Generalized matrix completion for predicting missing data entries with user constraints
Practical Uses
Auto-Fill
Fill missing data entries satisfying domain or user constraints. Indicates which constraints are important
Hyper-Spectral Images
Fully interpretable, end-member fingerprint identification and classification based on spectral fingerprints
Synergistic
Not aware of your problem constraints? Combine with HyperTune to generate a report of all possibilities