This is an experimental Project. The aim of this library is to autoselect visuaizations to evaluate machine learning models.
You can import th library locally i from this repo via
import vis_autselect.visualize as visualize
Instatiate a visualizer object:
vis = visualize.Visualizer()
You can import data in two different ways
- Pass in an output array from your ml experiment
vis.select(some_arr)
- Pass in an annotated array to help classify the input data
vis.input_dict({'Confidence Scores': y_score})
There is also the possibillity to import multiple object as such:
annotaded_data = {
'Confidence Scores': y_score,
'Confusion Matrix' : cm,
'ROC' : roc,
'Ground Truth Values' : y_test,
'Predictions' : y_pred
}
vis.input_dict(annotaded_data)
vis.info()
After you have inputed all the data you can just call the visualize()
function and it will generated all possible vissualizations.