Random Forest Machine Learning Tutorial in Python for Lithology Prediction - Includes Overview

Random forest is a very popular machine learning algorithm that can be used for both classification and regression. Within this tutorial we will go over the basics of the random forest algorithm before moving onto a real world example where we are attempting to predict a lithological class from well log measurements.
⭐️ If you haven't already, make sure you subscribe to the channel: / @andymcdonald42
Data Source
Bormann, Peter, Aursand, Peder, Dilib, Fahad, Manral, Surrender, & Dischington, Peter. (2020). FORCE 2020 Well well log and lithofacies dataset for machine learning competition [Data set]. Zenodo. doi.org/10.5281/zenodo.4351156
▼ --- SUPPORT THE CHANNEL --- ▼
☕️ BUY ME A COFFEE: www.buymeacoffee.com/andymcdo...
▼ --- GET THE CODE -- ▼
github.com/andymcdgeo/Petroph...
▼ --- RECOMMENDED BOOKS --- ▼
As an Amazon Associate I earn from qualifying purchases. By buying through any of the links below I will earn commission at no extra cost to you.
PYTHON FOR DATA ANALYSIS: Data Wrangling with Pandas, NumPy, and IPython
UK: amzn.to/3HNycJ9
US: amzn.to/3DL7qPv
FUNDAMENTALS OF PETROPHYSICS
UK: amzn.to/3l1PgSf
PETROPHYSICS: Theory and Practice of Measuring Reservoir Rock and Fluid Transport Properties
UK: amzn.to/30UNWZS
US: amzn.to/3DNqBbd
WELL LOGGING FOR EARTH SCIENTISTS
UK: amzn.to/3FHsbfn
US: amzn.to/3CILAuE
GEOLOGICAL INTERPRETATION OF WELL LOGS
UK: amzn.to/3l2v2HV
US: amzn.to/30UOTkU
▼ --- SOCIAL CHANNELS --- ▼
Thanks for watching, if you want to connect you can find me at the links below:
/ andymcdonaldgeo
/ geoandymcd
/ andymcdonaldgeo
www.andymcdonald.scot/
Be sure to sign up for my newsletter to be kept updated when I post and share new content on KZbin and Medium.
fabulous-founder-2965.ck.page...
#datascience #petrophysics #python #machinelearning #geosciences

[@AndyMcDonald42]

The notebook for this tutorial can be found here: github.com/andymcdgeo/Petrophysics-Python-Series/blob/master/27%20-%20Random%20Forest%20for%20Lithology%20Classification.ipynb

[@giopremiro9482]

As an IT Student looking forward to learning Data Science specifically Machine Learning, this is a great way to learn how to sort the data; clean it, verify its accuracy and present it. Ever since my university presented Machine Learning, I've been hooked ever since. I'm looking forward to watching more of your videos, please do keep uploading!

[@AndyMcDonald42]

Thanks Gio.

[@faicornelius2601]

Thanks for being a great teacher, Andy. Please can you do a video on Artificial neural networks in machine learning?

[@AndyMcDonald42]

Hi Fai, I have just finished an article version of using ANN to predict well log properties. I will hopefully be turning it into a video in the next few weeks. If you want, feel free to check out the article version here: towardsdatascience.com/how-to-create-a-simple-neural-network-model-in-python-70697967738f

[@Jean-tf4gh]

Hello Andy thanks for your excellent channel, I am trying to use this workflow for predicting facies, those faces exhibit a significant imbalance in the distribution, I mean some of them only have a few quantities, but others have extremely high amounts, so using train_test_split could no ensure to cover those facies with low presence, so could you please explain to us how to deal with this problem, I was reading about (StratifiedKFold, KFold) but I am no sure how to use it.

[@AndyMcDonald42]

Hi Jean. Dataset imbalance is something that I am looking into at the moment. Using k-fold validation is one way to combat it, however, the problem still exists where you may have only a few samples of one facies, and a massive amount of samples for another, such as shale. There are a few ways to deal with imbalanced datasets, which involve resampling, but I do not believe that is appropriate to geological data. It is definitely an area within geoscience/petrophysics that needs further research

[@tommasoseneca9189]

Hi there! Great channel! I loved this video, but I have a question: once we have a model and we have found that it is quite accurate (in your video 91%), is it possibile to put as input a row of values (our X) to make a prediction (y) of the specific rock? My idea is the following one: I have a new input line appending to my df, can I predict its y value (thus, the rock)? How can I do that? In a very basic form it should be something like this: prediction=Trained_Model(new_X_row)... Thanks!

[@AndyMcDonald42]

Thanks. Yes that is possible. Once the model has been trained, you can then use it to predict on new data (model.predict(X_values)) as long as you have the same input features. Let me know if you have any issues.

[@tommasoseneca9189]

@@AndyMcDonald42 Done! Great ✌🏻 Thanks a lot!

[@ravinayak2785]

Can you please send me the prediction of astroid orbit path using random forest algorithm project video

[@trolololo88]

hello! how to convert las files to csv in a proper way? any tutorials for that theme?

[@AndyMcDonald42]

I have a video planned to show this process. However, in the mean time, you can easily do this using the LASIO library, converting the data to a Pandas dataframe and then exporting to CSV. Hope this helps.

[@faisalkhalifa8554]

HI Andy thanks for the great work you are doing I am learning a lot from you . can you please check the link for the code in this video it seems it is the wrong one as it took me to the earthquake code

[@AndyMcDonald42]

Thanks Faisal. Sorry for that. I have updated the GitHub repo with my in progress files at github.com/andymcdgeo/Petrophysics-Python-Series You will find the Random Forest code in notebook 27. I will update the notebook properly soon with documentation.