Project Botticelli

Machine Learning at PASS SQLRally Nordic 2015

14 January 2015 · 279 views

Attend my pre-conference seminar!

PASS SQLRally LogoI am delighted to be one of the presenters at PASS Rally Nordic 2015, taking place in Copenhagen, 2–4 March 2015. I will present a full-day, intensive, pre-conference seminar on Advanced Analytics with Azure Machine Learning, SQL Data Mining, and R. I will also speak briefly on this subject during the main conference.

If you want to get up-to-speed with the new ML technologies to start doing predictive analytics on your own data, this seminar is your first such opportunity anywhere in Europe. Go ahead, register, and see you in Copenhagen!

Pre-conference seminar in more detail:

Data mining and machine learning are having their renaissance: after more than 40 years of academic research those powerful, algorithmic techniques are finally within the reach of anyone who can understand data. They are fast, inexpensive and, best of all, they find patterns and they can generate insight that business people are asking for! Rafal Lukawiecki will share his decade of hands-on data mining experience during this intensive, full-day seminar as he teaches you about the newest, cloud-based Microsoft analytical toolkit: Azure Machine Learning (Azure ML) and its very useful on-prem companion, the SQL Server Data Mining engine contained in SSAS, supplemented with gentle “I am really not a statistician” use of the open source R software. While you might catch a glimpse of Excel this day will not focus on it, as we will spend 90% of our time in SSDT, SSMS, R, and, of course, in ML Studio. However, before we can have fun exploring patterns and making predictions, you must understand how a somewhat unusual (for a relational guy or a girl) form of data preparation leads to finding interesting, potentially great predictive results. We will cover:

  • Introduction to predictive analytics, data mining, machine learning, and rudimentary descriptive statistics, as well as data and content types used in the process (this will be at level 200)
  • Data preparation, model building, and its very important validation, on-premise, using SQL Server Data Mining (level 300)
  • Model building and testing in the cloud, using Azure ML (level 250)
  • Use and deployment of both on-premise models (using T-SQL and DMX) and cloud-based models (using REST web services API) at level 300
  • A level 300-400+ look into the specifics of a few interesting and useful algorithms, including Regression and Classification Decision Trees, Random Forests and Jungles, Association Rules, Neural Networks and Logistic Regression, and a brief, level 400+ look at balancing model performance to user requirements by selecting the correct sensitivity vs specificity thresholds.

At the end of the day you will have learned the entire process of machine learning and data mining, focusing on the often missed out part of correct data preparation, which should enable you to start experimenting with your own data straight away. Although we cannot promise that, we will also do our best to make sure you are tired, as this will be a very intensive day, so please sign up if you like the idea of learning lots in a short span of time.