We're starting with a lightning talk by previous attendee Alex Kelly, "Industry 4.0 and Machine Learning".
The hands-on session is on "feature selection". When we use a machine learning algorithm, we have to decide how it's going to use our data. We may need to clean up the data, and we will need to choose which features to use - sometimes, leaving stuff out leads to better results.
Paul Brabban will explain the idea and walk through a simple example using an Azure Notebook, visualising a simple data set and showing some basic techniques for feature selection. The walkthrough will be in Python, but you're welcome to use whatever language you prefer.
Attendees are then be invited to have a go themselves, with any data set they like - the session will end with an invite to show-and-tell what you discovered so that we can all learn!
This is a beginner-friendly session - no prior knowledge of statistics or ML will be required.
Problem of the Month?
A month is too long to go between Machine Learning meetups! We're going to select a "Problem of the Month" that we can use to keep our learning and practice up between the meetups. Members are invited to suggest a problem!
What to bring?
You'll need a laptop to try out the hands-on session. You can either use software installed on your own computer, or you can sign up for an online notebook supporting common languages with Microsoft Azure (notebooks.azure.com), Kaggle (www.kaggle.com). We'll try and help if you're struggling with setup.
Refreshments?
The Showroom Cafe Bar can provide water for free, or the usual kind of food and drink if you want to pay for it.
Code of Conduct?
We follow the code of conduct described at berlincodeofconduct.org
Freebies?
We're an O'Reilly Partner, and as a member you can get a free 30 day trial with Safari Books Online at www.safaribooksonline.com. They also give us a 20% discount for the Strata Conference conferences.oreilly.com with code PCSMLM20.
Code of Conduct:
berlincodeofconduct.org
More details and tickets: www.meetup.com
Attending: 1 person.