Dealing With Missing Values in R

Novice and advanced R users from all professional groups.

• CHF 600.- for members of UZH/ETH and associated institutes
• CHF 800.- for alumni of UZH/ETH, members of other universities, the public sector and non-profit organizations
• CHF 1200.- for companies

In many applied settings the data is often incomplete, which makes data analysis a challenging task. There is, however, an abundant literature, as well as more than 150 R packages that address the issue of missing data. This workshop provides a clear overview of the different methods and strategies to handle missing values. The workshop will focuss on (a) the inferential framework that aims to estimate parameters (including their variance) in the presence of missing data, (b) matrix completion methods that aim to impute as well as possible, (c) recent developments in supervised learning in the presence of missing data.

Other topics include:

• Missing data mechanisms
• the EM-algorithm
• Multiple Imputation
• PCA with missing values/ Imputation with PCA
• Handling variables of different nature (quantitative, cartegorical, etc.)
• Prediction with missing values (random forest with missing values)

During the workshop, participants will learn about as well as practice using R packages such as: mice, missMDA, Amelia, etc. All of which can be found on the Rmisstatic https://rmisstastic.netlify.com/ plateform along with a dedicated task view, which aims at giving an overview of main references, contributors, and tutorials on data analysis in the presence of missing data.

Prerequisites

• Basic familiarity with R syntax
• Some experience with data analysis in R (loading data, fitting models, using the built-in graphics functions)
• Familiarity and some practical experience with regression modeling and PCA.

February 27-28, 2020

Registration