article thumbnail

Explaining black-box models using attribute importance, PDPs, and LIME

Domino Data Lab

The interest in interpretation of machine learning has been rapidly accelerating in the last decade. This can be attributed to the popularity that machine learning algorithms, and more specifically deep learning, has been gaining in various domains. PDPs for the bicycle count prediction model (Molnar, 2009).

Modeling 139
article thumbnail

ML internals: Synthetic Minority Oversampling (SMOTE) Technique

Domino Data Lab

Machine Learning algorithms often need to handle highly-imbalanced datasets. A weighted nearest neighbor algorithm for learning with symbolic features. Machine Learning, 57–78. UCI machine learning repository. The unreasonable effectiveness of data. Programs for machine learning.

article thumbnail

The Business Intelligence Market – What’s Old is New

In(tegrate) the Clouds

As the data visualization, big data, Hadoop, Spark and self-service hype gives way to IoT, AI and Machine Learning, I dug up an old parody post on the business intelligence market circa 2007-2009 when cloud analytics was just a disruptive idea. Ad hoc query, data mining, information I’m still not finding.