Machine Learning Workshop - April 2016

This time Pythonpune decided to do something new, so we organized Machine Learning Workshop. It was on 29 th April 2016 at Red Hat India Private Ltd. This meetup was one of the most successful meetup of Pythonpune in its history.

For this meetup 70 peoples were RSVPed. Around 70-80 peoples were present in the meetup. The response from people was tremendous. Most of the people were professionals who already knew Machine Learning, some of them were new to Machine Learning but they had Python knowledge and some students who knew Python with basics of Machine Learning.

First speaker in the meetup was Dr. Satish Patil, founder and chief data scientist of Lemoxo Technologies. He holds a PhD from the University of Minnesota, USA. Second speaker for the meetup - Sudarshan Gadhave who works for John Deere as Tech Lead.

Meetup started around 10.45 AM, Sudarshan introduced about Machine Learning, what is Machine Learning, why ML came in to the picture, etc. Then, he introduced Dr. Satish Patil, his education and his job profile.

Dr. Satish Patil is passionate about Maths, Design thinking and Cognitive Science. If you see his twitter handler (\@DataGeekSatish) it is also contain data and geek words ;) .

He introduced Machine Learning and its concepts, its applications in :

Self driving Google car.
Online recommendations such as Netflix and Amazon.
Fraud detection.
OLA Cabs.

Then he talked about Black box Machine Learning. Black box Machine Learning is like automatic programming. Automatic programming can give you a program that can solve a well defined problem. So you can imagine how complex it is.

Next topic was features. In ML and pattern recognition, a feature is an individual measurable property of a phenomenon being observed. Choosing informative and discriminating and independent features is a crucial step for effective algorithms in pattern recognition, classification, and regression. It is possible to select this features automatically in your data.

Then he introduced Training set, Test set, data set, classification, and clustering. I'll explain it in a brief.

A Training set is a set of data used to discover potentially predictive relationships.
A Test set is a set of data used to assess the strength and utility of a predictive relationship.
Classification is problem of identifying to which set of categories a new observation belongs.
Clustering is the assignments of set of observations in to subsets.

He also introduced other concepts, which are as follows:

Data pre-processiong
Market basket analysis
Entropy and Decision Tree.
Supervised and Unsupervised learning.
K-means algorithm.

Then he took some "hands-on" on Titanic Data Set. This data is about titanic survivals. Some peoples were able to do this "hands-on" session.

Machine Learning is very vast concept. At the end of talk Satish and Sudarshan advised to people who are new to ML is that:

If you are new to ML then start with Statistics (Math).
Then proceed to Data Mining.
Then slowly start with beginning of ML.

That will take a time but ML concepts are basis on that theory. So this is an useful advice to people like me :) who are new to ML.

After the Satish's talk, Chandan introduced us to Fedora Labs. It is a basic introduction to packages used for scientific purpose which includes C, C++ libraries for scientific development, Latex packages for scientific models. And how easily you can install it in your PC. For scientific help, he introduced #fedora-science IRC channel on Freenode.

Like I said above that this was the best meetup so far. We got positive feedback on meetup page and through private messages. Thank you Satish and Sudarshan for a wonderful talk on Introduction to Machine Learning. That was the best talk ever.

Thank you Red Hat India Pvt. Ltd. for the venue and Thanks Chandan and Praveen Kumanr for organizing meetup and choosing me as an event host.