A series of notebooks exploring different Machine Learning methods from Scikit-Learn.

• in Geoscience ML notebook 1, I look at methods for checking, tidying, and visualizing a dataset.

• in Geoscience ML notebook 2, I showcase some inferential statistics tools I built for selecting featues.

• in Geoscience_ML_notebook 3, I show how to use the prediction interval of a regression to discuss drilling economic risk.

• in Geoscience_ML_notebook 4, I take a quick look at what a simple neural network can learn about drilling problems due to mud loss.

• For Geoscience ML notebook 5 a visual tutorial on Support Vector Machines, check the ML repo, and read the full article here

• Geoscience ML notebook 6 is currently an evolving notebook. In it I evaluate and compare the performance of different ML methods in predicting a geophysical log from a suite of other logs; I will start with the compressional sonic initially, then see how well we can predict the shear sonic as compared with modeling. The log suite is from the same well (from Pev Avseth PhD Thesis) that Alessandro Amato del Monte used in the Seismic Petrophysics Notebook accompanying his Geophysical tutorial article on The Leading Edge.

• in Geoscience ML notebook 7, I will summarize my experience in team with with Mark Dahl at the SEG 2016 Machine Learning contest.

To wet your appetites, here's an example of compressional sonic log prediction using a cross-validated linear model, which will be the benchmark to evaluate the performance of other models in notebook 3, such as SVM and Random Forest:

and below is a heatmap with the pairwise Spearman correlation coefficient between the logs I will use: