Basic Intro to LASSO and Python Integration
Chuck will provide a quick intro to a new feature in Stata 16. Lasso is intended for prediction and selects covariates that are jointly correlated with the variables that belong in the best-approximating model. Said differently, lasso estimates the variables that belong in the model.
Stata's new lasso tools let you extract real features from mountains of data. With those features, you can:
- Predict outcomes
- Characterize groups and patterns in your data
- Search over highly nonlinear potential relationships
- Perform inference on covariates of interest
- Handle endogenous covariates or unobserved confounders
Power and Sample Size
In this talk Chuck introduces the concepts and jargon of power and sample size calculations such as alpha levels, power, and minimum detectable effect sizes. He does several simple calculations manually and then demonstrates how to replicate these calculations using Stata's -power- commands. Next Chuck demonstrates how to create tables and graphs for power, sample size, and minimum detectable effect sizes for a range of values. Lastly, he demonstrates how to calculate power using simulation methods and how to create your own custom power calculation programs that leverage Stata's -power- command to create custom tables and graphs.