Important Announcements:
  • Please Check Events Calendar for Upcoming Lunch Seminars Read more

  • Data Driven Drilling & Production Event Read more

  • Integrated Petrophysics for Reservoir Characterization Read more

  • STC - SPWLA Houston Chapter Read more

  • Flooding in Southeast Texas: The Science Behind the Floods Read more

Bootcamp: Machine Learning For Petrophysics


The three-day class will focus on how to apply machine learning techniques to petrophysical problems.
Students will become familiar with open source python toolboxes. You'll learn about the three main classes of machine learning. The main focus of the first day will be supervised and unsupervised learning. The afternoon will be spent stepping through the code and workflows, showing how to modify the code, and understand the outputs. For supervised learning, you will use logistic regression and support vector machines to pick formations. For unsupervised learning, you will use the K-means algorithm to observe how the formations form natural clusters in the data.
Over the next two days, we'll introduce additional algorithms and techniques: data wrangling (prepping data for analysis), feature engineering (adding information to the data using the data itself), and ensemble methods (machines that vote on the correct answer). The main goal would focus on how we increase accuracy in our formation predictions.
The students will walk away from the class with the actual code they can use to start playing around with problems they have at work. We will also introduce the students to the open source online community that is a huge resource when students and experts get stuck.
Laptop with administrator rights account. To be able to install necessary software (links will be sent to registered participants prior the course)
$350 for 3 days (including lunch and coffee breaks) [google registeration page with PayPal link]
October 23-25. 8am-4pm
University of Houston main campus (directions will be sent to registered participants)

For Registration - Click Here