FANS Network Meeting - Data Science Network | December 11, 2020 | Virtual/MSTeams

[DorteBrix]

Date: 11/12-2020

Time: 09:00-11:00 CET

Where: Microsoft Teams. Link and information will be sent with the confirmation mail, when you register for the meeting.

Welcome to this new Data Science Network that is cross Nordic and is held in English.

Focus is on knowledge sharing for people working with Data Science.

Agenda:

• Welcome
  • By: Frans Holm, SAS

• Data Science Use Cases
  • First I look on how to utilize Association Analysis for the purpose Data Exploration.
  • Then I am examine the Benford law and how you can use it to detect Fraud.
  • By: Jacob Mardfelt, SAS
  •  

• Ethical AI
  • AI is increasingly involved in decision making in contexts that can have a large impact on the individual.  Technology is today, to a large extent, ahead of regulations, making it necessary for us to look into ethical aspects to make sure that we deploy Responsible AI and avoid unwanted effects such as bias and discrimination. How can we do that and what do we need to have in place.
  • By: Josefin Rosén, SAS

• Introduction to Linear Mixed Models
  • Mixed models are used in a wide variety of fields, such as, social science, economics, business, and biomedical science.  In this presentation, we will introduce the concepts behind and demonstrate how you can use linear mixed models. We will look at how you can use both SAS and R to analyze a dataset with repeated measurements.
  • By: Ina Conrado, SAS

• Penalized regressions - where regression analysis meets machine learning
  • Ridge, Lasso and elasticnet regressions are becoming a type of linear regressions that can help in avoiding overfit, while performing model selection automatically. They are all based on an ordinary least squares estimation with a penalty term. This term varies between the types of penalized regressions. We will discuss the differences between the 3 main models, when to use which model and how they are implemented
  • By: Rune Hjorth Nielsen, SAS

• Closing & Lottery
  • By. Frans Holm, SAS

The purpose of the Nordic Data Science Network is to facilitate the knowledge sharing of data science in the broad interpretation of the term.

• We will for example discuss a specific estimation technique, model governance, implementation of models, the biases in model development and much more. The approach will be applied, with a focus on understanding a specific topic rather than a specific detail, and thereby enabling the audience to work further with the introduced topics.
• To increase the applied focus, we hope that companies will be interested in joining the network, by presenting their interesting projects to an audience of data scientist.
• While the network naturally will have a focus on SAS software, open source software will be welcome as well, as it fits with the open source connectivity of SAS.

SAS Nordic User Newsletter

Are you aware of our monthly SAS Nordic User Newsletter? Here you will find information about SAS Nordic user activities and events. Stay informed, subscribe here only by entering your e-mail. You can unsubscribe at any time.

FANS - Nordic SAS Users Group page