Vergangenes Meetup

Meetup #8: LIME | End-2-End Data Science

Dieses Meetup liegt in der Vergangenheit

28 Personen haben teilgenommen

Bild des Veranstaltungsortes

Details

It’s getting colder and thus high time for more hot data science topics in Münster Central. Let‘s get together to explore LIME together with Shirin. And we are also curious to hear Mark’s thoughts on the full data science circle. Read more below and RSVP before the limited seats are taken.

See you soon!

Due to high no-show rates at the last two meetups we want to discourage blocking our limited seats and then not showing up. From now on we will check who RSVP‘ed but doesn’t show up. We would then remove the no-show people from the group.

----------------------

Shirin Glander

"Explaining complex machine learning models with LIME"

Shirin (https://www.linkedin.com/in/shirin-glander-01120881/) is Data Scientist at codecentric, runs a beautiful blog (http://shiring.github.io/) and organizes the R User Group (https://www.meetup.com/Munster-R-Users-Group/) in Münster.

Machine Learning models are inherently difficult to explain because of their complexity. This makes it hard to understand specific decisions. Local Interpretable Model-Agnostic Explanations (LIME) is a method to explain the decisions made by a machine learning model. It works with the underlying assumption that we can locally fit linear models to the data, even if the global model is vastly more complex. Shirin will introduce the LIME approach and show examples for how to explain different models in R and Python.

----------------------

Mark Keinhörster

"Data Science End-2-End"

Mark (https://www.linkedin.com/in/mark-keinh%C3%B6rster-1632a2109/) is Data Plumber at codecentric in Münster. He enjoys putting data-science code to production. The boundary between software development, data science and classical data analysis blurs with every new technology, whether it is IPython, Jupyter Notebooks, Hadoop with Hive or Pig or Apache Spark. Hence data analytics can become a great opportunity or a bottleneck. Where is the boundary between development and business? Where does modeling and requirements engineering take place and how do these relate to the actual software engineering? Let's create a development environment that encourages business guys to play, Data Scientists to rapidly prototype and gives developers the opportunity to translate models with little effort into production ready code.