Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 11

AZURE ML

WHAT IS THIS AND HOW CAN I USE IT?


Warren Sifre

This document and the information contained herein is confidential and proprietary to Allegient LLC and shall not be duplicated, used or disclosed
in whole or in part for any purpose other than review. All trademarks and/or service marks contained within this document are the property of their
respective owners. Allegient does not in any way warrant the use of their products and/or services offerings.

 2016 Allegient. All rights reserved.


ABOUT ME
• Managing Consultant at Allegient.

• In the IT Industry since 1998.

• Developed system integration solutions against many different


database platforms.

• Passion in Solutions Architecture at both hardware and software


levels.

• Interests in SQL Server, MongoDB, Hadoop,


Python/C#/Java/Powershell and Information Security (Hacking)

 2016 Allegient. All rights reserved.


SESSION AGENDA

• What is Machine Learning?

• Techniques of Learning

• What is Azure ML?

• Demonstration

 2016 Allegient. All rights reserved.


WHAT IS MACHINE LEARNING?

 2016 Allegient. All rights reserved.


WHAT IS MACHINE LEARNING?
Formal definition: “A computer program is said to learn from experience E with respect to
some class of tasks T and performance measure P, if its performance at tasks in T, as
measured by P, improves with experience E” - Tom M. Mitchell

Another definition: “The goal of machine learning is to program computers to use example
data or past experience to solve a given problem.” – Introduction to Machine Learning, 2 Edition, MIT Press
nd

“In general, ML converts data sets into pieces of software,


known as ‘models’, that can represent the data set and
generalize to make predictions on new data.” - John Platt
http://blogs.technet.com/b/machinelearning/archive/2014/07/01/what-is-machine-learning.aspx

 2016 Allegient. All rights reserved.


TECHNIQUES OF LEARNING
• Supervised
o Learning by using a set number of inputs and outputs.
o Limited in providing deep undiscovered insights.
o Used when you want to predict unknown answers from
answers you already have. (Historical Data)

• Unsupervised
o Learning by using all inputs.
o Can learn larger and more complex models.
o Can take longer to learn depending on the number of
linear levels.
o Used when you want to find unknown answers directly
from data. (No simple way to validate results.)

 2016 Allegient. All rights reserved.


WHAT IS AZURE ML?
• A cloud-based platform for designing, developing, testing and
deploying predictive models.

• There are three conceptual areas Experiments, Web Service and


ML Studio.
o Experiments –Predictive models for both training and scoring.
o Web Service – Data Interaction point for predictive models.
o ML Studio – Interface for developing and training predictive models.

• Leverage existing R, Python, OpenCV and other industry adopted


predictive algorithms.

 2016 Allegient. All rights reserved.


ALGORITHM SELECTION???

http://azure.microsoft.com/en-us/documentation/articles/machine-learning-algorithm-cheat-sheet/

 2016 Allegient. All rights reserved.


DEMONSTRATION

 2016 Allegient. All rights reserved.


HELPFUL LINKS

• Azure ML Studio
o https://studio.azureml.net/

• Azure ML – Microsoft Virtual Academy


o http://
www.microsoftvirtualacademy.com/training-courses/getting-started-with-micros
oft-azure-machine-learning

• Azure ML Algorithm Cheat Sheet


o http://azure.microsoft.com/en-us/documentation/articles/machine-learning-al
gorithm-cheat-sheet
/

 2016 Allegient. All rights reserved.


Warren Sifre
Email: wsifre@allegient.com
Twitter: @WAS_SQL
LinkedIn: www.linkedin.com/in/wsifre

 2016 Allegient. All rights reserved.

You might also like