Azure Machine Learning Algorithm Accuracy Enhancement, Tips and Tricks 7/20/2018 11:17 PM THR3078 Azure Machine Learning Algorithm Accuracy Enhancement, Tips and Tricks Leila Etaati AI MVP, PhD, Consultant © Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

Leila Etaati Microsoft AI MVP, PhD, Senior Consultant, Trainer and Data Scientist. International speaker in Microsoft Ignite USA 2017, Microsoft Insight Summit 2017, PASS Summit 2017, Microsoft NZ Ignite 2016, PASS BA, PASS24H, SQLRally, SQL Saturday in Oregon, Vienna, Auckland, Melbourne, Sydney, Brisbane. http://radacad.com/author/leila

Azure ML

Machine Learning Process What is Business Problem Collecting & Cleaning Data Choose Model &Parameters Evaluate Model Test Model Train Model

What is Business Problem

What is Business Problem Prediction Predictive Analytics Grouping Descriptive Analysis Find Unusual Data Point Descriptive-Anomaly Detection

What is Business Problem -Predictive Analytics Predict a Group Predict a Value

What is Business Problem - Descriptive Analytics

What is Business Problem - Anomaly Detection

Collecting & Cleaning Data

Collecting & Cleaning Data-Clean Data 7/20/2018 11:17 PM Length Weight height area 2000 20k 1.1 k 0.45 Bring Data In one Scale Length Weight height area 0.2 0.4 0.67 0.45 Remove Outliers 5,2 22,25,28,25,27,30,31,22 60,70 Gender Female Male Female 1 Male 1 Convert to Indicator Values © Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

Collecting & Cleaning Data-Clean Data

Collecting & Cleaning Data- Feature Selection Gender Age No Children Charges Region BMI Smoker

Demo-Feature Selection 7/20/2018 11:17 PM Demo-Feature Selection Leila Etaati © Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

Collecting & Cleaning Data- Cross Validation 7/20/2018 11:17 PM Collecting & Cleaning Data- Cross Validation Data Data Data Data Data Data for Test Data for Train Test Train Test Test Test Test Apply Model Apply Model Apply Model Apply Model © Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

Demo-Cross Validation 7/20/2018 11:17 PM Demo-Cross Validation Leila Etaati © Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

Choose Model & Parameters

Choose Model & Parameters: Data Linearity

Two main problem: Over fit and Under fit

Two main problem: Over fit and Under fit

Choose Model (Parameters):Parameters Tuning Decision Tree, Forrest, .. Depth of tree, Number of Decision Tree Clustering- K-mean, Number of Clusters

Demo-Tune Model Hyperparameters 7/20/2018 11:17 PM Demo-Tune Model Hyperparameters Leila Etaati © Microsoft Corporation. All rights reserved. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

Machine Learning Process What is Business Problem Collecting & Cleaning Data Choose Model &Parameters Evaluate Model Test Model Train Model

Please evaluate this session Tech Ready 15 7/20/2018 Please evaluate this session From your Please expand notes window at bottom of slide and read. Then Delete this text box. PC or tablet: visit MyIgnite https://myignite.microsoft.com/evaluations Phone: download and use the Microsoft Ignite mobile app https://aka.ms/ignite.mobileapp Your input is important! © 2012 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.