Book Image

AWS SageMaker, Machine Learning and AI with Python [Video]

By : Chandra Lingam
Book Image

AWS SageMaker, Machine Learning and AI with Python [Video]

By: Chandra Lingam

Overview of this book

<p>This course is designed to make you an expert in AWS machine learning and it teaches you how to convert your cool ideas into highly scalable products in a matter of days. The biggest challenge for a data science professional is how to convert the proof-of-concept models into actual products that your customers can use. There are several courses on machine learning that teach you how to build models in R, Python, Matlab and so forth. However, converting a model into a scalable solution and integrating with your existing application requires a lot of effort and development. The real success of your ideas and concepts depends on how soon you can put the capabilities in the hands of your customers. With the AWS Machine Learning service, you can easily conduct experiments and test your concepts. Once you are happy, you can instantly scale them to support millions of requests. No separate development work is needed.</p> <h2>Style and Approach</h2> <p>This course is completely hands-on with examples using: AWS Web Console, Python Notebook Files, and Web clients built on AngularJS. You will also learn and integrate security into exercises using a variety of AWS provided capabilities including Cognito.</p>

Related Content you might be interested in

Current Title:

Small book image
AWS SageMaker, Machine Learning and AI with Python [Video]
Chandra Lingam
Jul, 2018 | 12 hours 58 minutes
No titles found
Table of Contents (24 chapters)
Free Chapter
1
Introduction
Introduction
Root Account Setup and Billing Dashboard Overview
Enable Access to Billing Data for IAM Users
Create Users Required For the Course
AWS Command Line Interface Tool Setup and Summary
Six Advantages of Cloud Computing
AWS Global Infrastructure Overview
2
2019 SageMaker House Keeping
Demo 1: S3 Bucket Setup
Demo 2: Setup Notebook Instance on SageMaker
Demo 3: Source Code and Data Setup - Git Clone (New Lecture)
3
2019 Machine Learning Concepts
Machine Learning Benefits, Concepts, Terminologies
Data Types - How to handle mixed data types
Demo 4: Introduction to Python Notebook environment
Demo 5: Working with Missing Data
Demo 6: Data Visualization - Linear, Log, Quadratic and More
4
SageMaker Service Introduction
SageMaker Overview
Compute Instance Families and Pricing
Algorithms and Data Formats Supported For Training and Inference
5
A.2 XGBoost
XGBoost - Introduction and Comparison with Other Approaches
Source Code Overview
Demo 4: Create Files in SageMaker Data Formats and Save Files to S3
Demo 5: Working with XGBoost - Linear Regression Straight Line Fit
Demo 6: XGBoost Example with Quadratic Fit
Demo 7: Kaggle Bike Rental Data Setup, Exploration and Preparation
Demo 8: Kaggle Bike Rental Model Version 1
Demo 9: Kaggle Bike Rental Model Version 2
Demo 10: Kaggle Bike Rental Model Version 3
Demo 11: Training on SageMaker Cloud - Kaggle Bike Rental Model Version 3
Demo 12: Invoking SageMaker Model Endpoints for Real Time Predictions
Demo 13: Invoking SageMaker Model Endpoints from Client Outside of AWS
XGBoost Hyper Parameter Tuning
Demo 14: XGBoost Multi-Class Classification Iris Data
Demo 15: XGBoost Binary Classifier for Diabetes Prediction
Demo 16: XGBoost Binary Classifier for Edible Mushroom Prediction
Summary – XGBoost
6
A.3. SageMaker - Principal Component Analysis (PCA)
Introduction PCA and SageMaker PCA
PCA Demo Overview
Demo 1: PCA with Random Dataset
Demo 2: PCA with Correlated Dataset
Demo 3.1: PCA with Kaggle Bike Sharing - Overview and Normalization
Demo 3.2: PCA Local Model with Kaggle Bike Train
Demo 3.3: PCA training with SageMaker
Demo 3.4: PCA Projection with SageMaker
Summary
7
A.4. Factorization Machines For Recommender Systems and Click Rate Prediction
Introduction to Factorization Machines
Demo - Movie Recommender Data Preparation
Demo - Movie Recommender Model Training
Demo - Movie Predictions by User
8
DeepAR Time Series Forecasting
Introduction to DeepAR Time Series Forecasting
DeepAR Training and Inference Formats
Working with Time Series Data, Handling Missing Values
Demo - Bike Rental as Time Series Forecasting Problem
Demo - Bike Rental Model Training
Demo - Bike Rental Prediction
Demo - DeepAR Categories
Demo - DeepAR Dynamic Features Data Preparation
Demo - DeepAR Dynamic Features Training and Prediction
Summary
9
AWS Machine Learning Service
Python Development Environment and Boto3 Setup
Project Source Code and Data Setup
Lab: Intro to Python Jupyter Notebook Environment, Pandas, Matplotlib
Lab: AWS S3 Bucket Setup and Configure Security
Summary
Machine Learning Terminology
Data Types supported by AWS Machine Learning
Linear Regression Introduction
Binary Classification Introduction
Multiclass Classification Introduction
Data Visualization - Linear, Log, Quadratic and More
10
Linear Regression
Lab: Linear Model, Squared Error Loss Function, Stochastic Gradient Descent
Lab: Linear Regression for complex shapes
Summary
11
AWS - Linear Regression Models
Lab: Simple Training Data
Lab: Datasource
Lab: Train Model with default recipe
Concept - How to evaluate regression model accuracy?
Lab: Evaluate predictive quality of the trained model
Lab: Review Default Recipe Settings Used to Train model
Lab: Train Model with Custom Recipe and Review Performance
Model Performance Summary and Conclusion
12
Adding Features to Improve Model
Lab: Quadratic Fit Training Data
Lab: Under fitting with Linear Features
Lab: Normal Fit with Quadratic Features
Summary
13
Normalization
Lab: Impact of Features with Different Magnitude
Concept: Normalization to smoothen magnitude differences
Lab: Train Model with Feature Normalization
Summary
14
Adding Complex Features
Lab: Prepare Training Data
Lab: Adding Complex Features
Lab: Train Model with Higher Order Features
Lab: Performance of Model with Degree 1 Features
Lab: Performance of Model with Degree 4 Features
Lab: Performance of Model with Degree 15 Features
Summary
15
Kaggle Bike Hourly Rental Prediction
Review Kaggle Bike Train Problem and Dataset
Lab: Train Model to Predict Hourly Rental
Lab: Evaluate Prediction Quality
Linear Regression Wrap up and Summary
16
Logistic Regression
Binary Classification - Logistic Regression, Loss Function, Optimization
Lab: Binary Classification Approach
True Positive, True Negative, False Positive and False Negative
Lab: Logistic Optimization Objectives
Lab: Logistic Cost Function
Lab: Cost Example
Optimizing Weights
Summary
17
Onset of Diabetes Prediction
Problem Objective, Input Data and Strategy
Lab: Prepare For Training
Lab: Training a Classification Model
Concept: Classification Metrics
Concept: Classification Insights with AWS Histograms
Concept: AUC Metric
Lab: Review Diabetes Model Performance
Lab: Cutoff Threshold Interactive Testing
Lab: Evaluating Prediction Quality with Additional Dataset
Lab: Batch Prediction and Compute Metrics
Summary
18
Multiclass Classifiers using Multinomial Logistic Regression
Lab: Iris Classification
Lab: Train Classifier with Default and Custom Recipe
Concept: Evaluating Predictive Quality of Multiclass Classifiers
Concept: Confusion Matrix to Evaluating Predictive Quality
Lab: Evaluate Performance of Iris Classifiers using Default Recipe
Lab: Evaluate Performance of Iris Classifiers using Custom Recipe
Lab: Batch Prediction and Computing Metrics using Python Code
Summary
19
Text Based Classification with AWS Twitter Dataset
AWS Twitter Feed Classification for Customer Service
Lab: Train, Evaluate Model and Assess Predictive Quality
Lab: Interactive Prediction with AWS
Logistic Regression Summary
20
Data Transformation using Recipes
Recipe Overview
Recipe Example
Text Transformation
Numeric Transformation - Quantile Binning
Numeric Transformation – Normalization
Cartesian product Transformation - Categorical and Text
Summary
21
Hyper Parameters, Model Optimization and Lifecycle
Introduction
Data Rearrangement, Maximum Model Size, Passes, Shuffle Type
Regularization, Learning Rate
Regularization Effect
Improving Model Quality
Model Maintenance
AWS Machine Learning System Limits
AWS Machine Learning Pricing
22
Integration of AWS Machine Learning With Your Application
Introduction
Integration Scenarios
Security using IAM
Hands-on lab - List of Demos and Objective
Lab: Enable Real Time End Point and Configure IAM Prediction User
Lab: Invoking Prediction from AWS Command Line Interface
Lab: Invoking Prediction from Python Client
Lab: Python Client to Train, Evaluate Models and Integrate with AWS
23
Web Client with Cognito and AngularJS
Lab: Invoking Prediction from Web Page AngularJS Client
Demo Allowing Prediction Only For Registered Users
Cognito Overview
Lab: Cognito User Pool Configuration
Lab: AngularJS Web Client - Invoke Prediction for authorized users
Lab: Invoke Machine Learning Service From AWS EC2 Instance
Summary
24
Conclusion
Conclusion
Chapter 20
Data Transformation using Recipes
Content Locked
Section 5
Numeric Transformation – Normalization
Learn about normalization transformation in this video.
Previous Section
End of Section 5
Next Section