A simplified Integrated Assesment Model calibration and projection.
A hybrid between unsupervised and supervised learning that improves topic modelling by including guiding external information. In this post I go over Blei and Mcauliffe (2008) model with comments and derivations.
For those that know very well regressions but not so much about Tensor Flow, it is always a good idea to learn in a familiar enviroment.
This post shows how to build a deep neural network from scratch in Julia for the classic appication of building a clasifier of images of cats
LDA has become an industry standard for topic extraction, this post revisit the math behind the model, from words to numbers.