»  Home »  Training »  Classroom and web-based training courses »  Lasso Using Stata: Methods for Prediction and Inference

## Lasso Using Stata: Methods for Prediction and Inference

Description

Learn how to use lasso to predict outcomes and estimate effects when you are faced with many variables–perhaps poorly described and poorly understood variables–and classical regression techniques break down. This course provides introductions to lasso in Stata and to using lasso-based estimators in Stata for causal parameters in a high-dimensional model.

The first part of the course provides introductions to lasso and to elastic net for linear-outcome models and for nonlinear-outcome models. It also demonstrates how to use Stata to predict these outcomes.

The second part of the course provides introductions to high-dimensional models and to estimation methods for causal parameters in these models. The introductions show how the lasso is used in these methods. High-dimensional models for linear models, nonlinear models, and linear models with endogenous variables are discussed.

Stata examples based on real data and simulated data are used to illustrate the methods.

Price: \$1,295

We offer a 15% discount for group enrollments of three or more participants.

David M. Drukker is the Executive Director of Econometrics at Stata. His passion for programming enabled him to start working at Stata in 2000, when he finished his Ph.D. in economics at the University of Texas at Austin. Since then he has developed many Stata commands for estimating treatment effects and for analyzing panel data, time-series data, and cross-sectional data. He played a key role in the initial development of Stata MP, helped integrate Mata into Stata, and helped develop Stata and numerical techniques.

Course topics

• Lasso for prediction
• Methods: lasso and elastic net
• Linear models for continuous outcomes
• Logit models for binary outcomes
• Poisson models for count outcomes
• Selection of variables via cross-validation, adaptive lasso, and plugin estimators
• Out-of-sample prediction
• Lasso for inference in high-dimensional models
• Estimation of coefficients, odds ratios, and incidence-rate ratios with standard errors and confidence intervals for variables of interest
• Using lasso methods to select from many potential control variables
• Models: linear, logit, Poisson, and linear with endogenous covariates
• Methods: partialing out, double selection, and cross-fit partialing out

Prerequisites

A general familiarity with Stata and a graduate-level course in regression analysis or comparable experience.

Next session

Currently, there are no scheduled sessions of this course.

Inform me of the next offering of Lasso Using Stata: Methods for Prediction and Inference.

Name:

Email:

Country:

Do you want to be notified of all upcoming training opportunities? Sign up for our convenient email alerts.

Notes

Enrollment is limited. This course is offered in both classroom and web-based settings.

Classroom training courses are two-day courses that run from 8:30 a.m. to 4:30 p.m. each day. These courses take place at a training center where computers with Stata installed are provided. A continental breakfast, lunch, and an afternoon snack will also be provided; the breakfast is available before the course begins.

Web-based training courses are four-day courses that run for three and a half hours each day. You will be provided with a temporary Stata license to install on your computer, a printed copy of the course notes, and all the course datasets so that you can easily follow along. Learn more about how our web-based training courses work, watch a video demonstration, and find technical requirements for participating in this type of training.