AutoFeat computes the relationships between N tables from the {dataset} repository, using {coma, jaccard} similarity score with a threshold of X (i.e., all the relationships with a similarity < threshold will be discarded).
AutoFeat creates M join trees: the best performing join tree is {ID_number}
< print paths with the data quality score >
< print the selected features with the relevance and redundancy score >
< print ML result: accuracy, model, feature importance >
Augment dataset has flag explain.