Metric learning-based predictive model for small, high dimensional datasets
A. Orlichenko et al., "Latent Similarity Identifies Important Functional Connections for Phenotype Prediction," in IEEE Transactions on Biomedical Engineering, doi: 10.1109/TBME.2022.3232964.
- Very fast runtime
- High accuracy on limited data
- Multimodal
- Sklearn interface
- python
- pytorch with cuda
- numpy
- sklearn
- requests (to get sample data)
Take a look at the example in the notebooks directory for sample usage.
from sklearn.model_selection import train_test_split
from latsim import LatSimClf
...
xtr, xt, ytr, yt = train_test_split(x, y, stratify=y, train_size=0.75)
clf = LatSimClf().fit(xtr,ytr,ld=1)
yhat = clf.predict(xt)
An interactive demo was available here. We are working to put up another version.
Anton Orlichenko | aorlichenko@tulane.edu
aorliche.github.io
MBB Laboratory