|
|
 |
Listserv Message Center
 |
Machine Learning for Prediction and Causal Inference |
 |
|
|
|
| Posted by: | Michael Zyphur |
| Title/Position: | Director |
| School/Organization: | Institute for Statistical and Data Science |
| Sent to listserv of: | SESP, SPSSI |
| Date posted: | October 25th, 2022 |
|
|
|
|
 |
 |
|
Hi everyone,
Instats presents Machine Learning for Prediction and Causal Inference, a 3-day seminar taught by University of Cambridge professor of economics Melvyn Weeks, running December 14-16. Professor Weeks has extensive experience teaching and using machine learning (ML) methods for applied social and health science problems, and his introductory seminar assumes no prior knowledge beyond basic regression analysis in R, Stata, or Python -- he'll be covering the applications with worked examples in all three packages!
https://instats.com/seminar/machine-learning-for-prediction-and-caus6271
In this 3-day seminar, you'll be introduced to the logic of ML methods and shown how to apply them to address practical problems of prediction and causal inference. An emphasis will be placed on using ML to uncover heterogeneity in causal effects across the members of a population. For example, researchers may want to infer the effect of an economic, educational, or public health intervention, or a firm may seek to understand how a change in pricing will impact aggregate demand. In these cases, the interest may be in an average effect, but also how the effect varies over different segments of the population (i.e., heterogeneity in the effect). This seminar will provide you with the tools to undertake such inquiry using machine learning (ML), while ensuring that you understand and can communicate how the methods work for prediction and causal inference. An official Instats certificate of completion is provided at the conclusion of the seminar. For European PhD students, the seminar offers 2 ECTS equivalent points.
By the end of this two-day seminar you will be able to:
--Understand the relationship between regression and ML methods;
--Distinguish different ML methods including supervised and unsupervised;
--Run different types of ML approaches for prediction in R, Stata, or Python;
--Evaluate causal effects using ML;
--Account for an model heterogeneity in these effects; and
--Describe ML methods for scientific applications and writing up results.
The 2-day livestreamed seminar will be held over Zoom, but you can participate asynchronously through the course webpage where all videos will be posted immediately after each day. The seminar webpage also offers 30 days of Q&A support by professor Weeks, and a certificate of completion is provided at the conclusion of the seminar. Furthermore, as with all Instats seminars, a substantial discount is offered to PhD students and others enrolled in a university degree, and the seminar offers 2 ECTS equivalent points for european PhD students.
Please feel free to share this widely with your colleagues, PhD students, and any others who you think might be interested -- and of course you can email me with any questions.
Happy modeling!
Mike
Michael Zyphur
Director
Institute for Statistical and Data Science
instats.com
|
|
 |
|
|
|
|
|
|
Forum Home Page
