Related Posts
Additional Posts in Data & Analytics Consultants
Thought this was interesting. Across 160 teams of researchers, just about all failed to make good life outcome predictions on things like GPA, evictions, layoffs, and others. Data followed 4.5k families across 15 years, with 13k features (varied over time). Haven't looked at it directly yet, but will be turning the docs and data inside out... In the meantime, authors claim this as showing the limits of ML. Oh, and it's published in PNAS, so you know there's some big publication energy there.
https://www.pnas.org/content/117/15/8398
What is a data lake in basic terms?
Has anyone else begun to resent data science?
New to Fishbowl?
unlock all discussions on Fishbowl.



Compare it with a baseline.
Baseline can be: randomly chosen or a rule-based approach or a supervised approach.
Underpromise, overdeliver
Lol it took me a while to get what C1 was saying…
But OP, did you mean: no supervisor OR unsupervised learning?
Apologies for the lack of clarity. I meant unsupervised learning!
Classification metrics if you're trying these algos as just something else to try on a classification task. But that's very very unreliable, so... Don't.
Otherwise, look at any number of algo appropriate metrics that are basically what makes the algo work in the first place. Apriori has support and confidence, clustering has a huge number of cluster distance and similarity metrics, PCA and friends have their own stuff with variance in the eigenvectors, information theoretic methods have their own similarity metrics they use, autoencoders use whatever similarity metrics you opt for, etc etc....
Thank you for the detailed reply, I’ll use these as a starting point!