Date: 23/9-2021
Time: 09:30-12:30 CET (GMT +1)
Places:
- Denmark: Virtual
- Finland: Virtual
- Norway: Virtual
- Sweden: Virtual
- Virtual: Microsoft Teams (join link will be sent with the registration confirmation email)
Agenda
- How to create simulated data (distributions)
- Generate data sets with specific distributions
- Generate variables with dependencies in data
- Generate time series datasets
- By Jussi Varjus, SAS
- The use of simulated data to ensure that your models are "correct" or ethical.
- It can be cumbersome to ensure that analytical / machine learning models are "correct" or ethical. There are several examples of models reproducing historical bias or confirming old technical principles. The presentation will show some ideas to try and see if you can find biases using simulated data.
- By Pål Navestad, ConocoPhillips
- What was the effect of the pandemic on various Danish timeseries? - an analysis using SAS
- Focus is on usable results, but also the choice of model
- By Professor Anders Milhøj, University of Copenhagen
- Communication of Machine Learning results!
- Presentation and discussions about why it is so difficult.
- By Sara Armandi, SAS
Welcome!
Subscribe to Nordic SAS User Newsletter (monthly)
Back to SAS Nordic User site
EthicalAISimulatedFunWageData.zip
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.