book

Smarter Decisions - The Intersection of Internet of Things and Decision Science

Name: Smarter Decisions - The Intersection of Internet of Things and Decision Science
Author: Jojo Moolayil
ISBN: 9781785884191

by Jojo Moolayil

July 2016

Intermediate to advanced

392 pages

9h 38m

English

Packt Publishing

Read now

Unlock full access

Smarter Decisions – The Intersection of Internet of Things and Decision Science
Smarter Decisions – The Intersection of Internet of Things and Decision Science
Credits
About the Author
About the Reviewer
eBooks, discount offers, and more
Why subscribe?
Preface
What this book covers
What you need for this book
Who this book is for
Sections
Getting readyHow to do it…How it works…There's more…See also

Conventions
Reader feedback
Customer support
Downloading the example codeErrataPiracyQuestions
1. IoT and Decision Science
Understanding the IoTIoT in a real-life scenario
Demystifying M2M, IoT, IIoT, and IoE
Digging deeper into the logical stack of IoT
PeopleProcessesTechnologySoftwareProtocolInfrastructureBusiness processesThingsData
The problem life cycle
The problem landscape
The art of problem solving
The interdisciplinary approachThe problem universe
The problem solving framework
Summary
2. Studying the IoT Problem Universe and Designing a Use Case
Connected assets & connected operationsThe journey of connected things to smart thingsConnected assets - A real life scenarioConnected operations – The next revolutionWhat is Industry 4.0?
Defining the business use case
Defining the problemResearching and gathering contextGathering context - examining the type of problemGathering context - research and gather contextResearch outcomeHow is detergent manufactured?What are the common issues that arise in the detergent manufacturing process?What kind of machinery is used for the detergent manufacturing process?What do we need to know more about the company, its production environment, and operations?Prioritize and structure hypotheses based on the availability of dataValidating and Improving the hypotheses (iterate over #2 and #3)Assimilate results and render the story
Sensing the associated latent problems
Designing the heuristic driven hypotheses matrix (HDH)
Summary
3. The What and Why - Using Exploratory Decision Science for IoT
Identifying gold mines in data for decision makingExamining data sources for the hypothesesData surfacing for problem solvingEnd product related informationManufacturing environment informationRaw material dataOperational dataSummarizing the data surfacing activityFeature explorationUnderstanding the data landscapeDomain context for the data
Exploring each dimension of the IoT Ecosystem through data (Univariates)
What does the data say?Exploring Previous Product...Summarizing this section
Studying relationships
So what is correlation?Exploring Stage 1 dimensionsRevisiting the DDH matrix
Exploratory data analysis
So how do we validate our findings?So how does hypothesis testing work?Validating hypotheses - category 1How does the chi-squared test work in a nutshell?Validating hypotheses - category 2What does a Type 1 error mean?So what is ANOVA?Validating hypotheses - category 3So what is regression?Hypotheses - category 3Summarizing Exploratory Data Analysis phase
Root Cause Analysis
Synthesizing resultsVisualizing insightsStitching the Story togetherConclusionProduction QuantityRaw material quality parametersResources/Machinery used in Stage 3Assembly Line
Summary
4. Experimenting Predictive Analytics for IoT
Resurfacing the problem - What's next?
Linear regression - predicting a continuous outcome
PreludeSolving the prediction problemSo what is linear regression?Interpreting the regression outputsF statisticEstimate/coefficientsStandard error, t-value, and p valueResiduals, multiple R squared, residual standard error and adjusted R squaredWhat is the adjusted R-squared value?Improving the predictive modelLet's define our approachHow will we go about it?Let's being modelingSo how do we move ahead?The important points to ponder are as follows:What should we take care of?So what next?
Decision trees
Understanding decision treesSo what is a decision tree?How does a decision tree work?What are different types of decision trees?So how is a decision tree built and how does it work?How to select the root node?How are the decision nodes ordered/chosen?How different is the process for classification and regression?Predictive modeling with decision treesSo how do we approach?So what do we do to improve the results?So, what next? Do we try another modeling technique that could give us more powerful results?
Logistic Regression - Predicting a categorical outcome
So what is logistic regression?So how does the logistic regression work?How do we assess the goodness of fit or accuracy of the model?Too many new terms?Recap to the model interpretationImproving the classification modelLet's define our approachHow do we go about it?Let's begin modelingSo how do we move ahead?Adding interaction termsWhat can be done to improve this?What just happened?What can be done to improve the TNR and overall accuracy while keeping the TPR intact?
Summary
5. Enhancing Predictive Analytics with Machine Learning for IoT
A Brief Introduction to Machine LearningWhat exactly is ensemble modeling?Why should we choose ensemble models?So how does an ensemble model actually work?What are the different ensemble learning techniques?Quick Recap - Where were we previously?
Ensemble modeling - random forest
What is random forest?How do we build random forests in R?What are these new parameters?MtryBuilding a more tuned version of the random forest modelHow?Can we improve this further?What can we do to achieve this?
Ensemble modeling - XGBoost
What is different in XgBoost?Are we really getting good results?What next?A cautionary note
Neural Networks and Deep Learning
So what is so cool about neural networks and deep learning?What is a neural network?So what is deep learning?So what problems can neural networks and deep learning solve?So how does a neural network work?NeuronsEdgesActivation functionLearningSo what are the different types of neural networks?How do we go about modeling using a neural network or deep learning technique?What next?What have we achieved till now?
Packaging our results
A quick recapResults from our predictive modeling exerciseFew points to note
Summary
6. Fast track Decision Science with IoT
Setting context for the problemThe real problemWhat next?
Defining the problem and designing the approach
Building the SCQ: Situation - Complication - QuestionResearchHow does a solar panel ecosystem work?FunctioningWhat are the different kinds of solar panel installations?What challenges are faced in operations supported by solar panels?Domain contextDesigning the approachStudying the data landscape
Exploratory Data Analysis and Feature Engineering
So how does the consumption fare in comparison with the generation?BatteryLoadInverterAssimilate learnings from the data exploration exerciseLet's assimilate all our findings and learnings in briefSolving the problemFeature engineering
Building predictive model for the use case
Building a random forest model
Packaging the solution
Summary
7. Prescriptive Science and Decision Making
Using a layered approach and test control methods to outlive business disastersWhat is prescriptive analytics?What happened?Why and how did it happen?When will it happen (again)?So what, now what?Solving a prescriptive analytics use caseContext for the use caseDescriptive analytics - what happened?Inquisitive analytics - why and how did it happen?Predictive analytics – when will it happen?The inception of prescriptive analyticsGetting deeper with prescriptive analyticsSolving the use case the prescriptive wayTest and control analysisImplementing Test & Control Analysis in Prescriptive AnalyticsImproving IVR operations to increase the call completion rateReducing the repeat callsStaff training for increasing first call resolution rateTying back results to data-driven and heuristic-driven hypotheses
Connecting the dots in the problem universe
Story boarding - Making sense of the interconnected problems in the problem universe
Step 1 - ImmediateStep 2 - Future
Implementing the solution
Summary
8. Disruptions in IoT
Edge/fog computingExploring the fog computing model
Cognitive Computing - Disrupting intelligence from unstructured data
So how does cognitive computing work?Where do we see the use of cognitive computing?The storyThe bigger question is, how does all of this happen?
Next generation robotics and genomics
Robotics A bright future with IoT, Machine Learning, Edge & Cognitive ComputingGenomicsSo how does genomics relate to IoT?
Autonomous cars
Vision and inspirationSo how does an autonomous car work?Wait, what are we missing?Vehicle - to - environmentVehicle - to - vehicleVehicle - to - infrastructureThe future of autonomous cars
Privacy and security in IoT
VulnerabilityIntegrityPrivacySoftware infrastructureHardware infrastructureThe protocol infrastructure
Summary
9. A Promising Future with IoT
The IoT Business model - Asset or Device as a ServiceThe motivationReal life use case for Asset as a Service modelHow does it help business?Best case scenarioWorst case scenarioNeutral caseConclusionLeveraging Decision Science to empower the Asset as a Service model
Smartwatch – A booster to Healthcare IoT
Decision science in health dataConclusion
Smart healthcare - Connected Humans to Smart Humans
Evolving from connected cars to smart cars
Smart refuel assistantPredictive maintenanceAutonomous transportConcluding thoughts
Summary

Content preview from Smarter Decisions - The Intersection of Internet of Things and Decision Science

Ensemble modeling - random forest

Random forest is an extremely popular machine learning technique that is used mainly for classification and regression. As the algorithm builds multiple decision trees, we have already covered a substantial part of the foundation required for random forest. Let's quickly understand the algorithm and solve our previous problem better.

What is random forest?

Random forest is a machine learning technique built on the principle of ensemble modeling. It builds an ensemble of decision trees with each tree having a randomly chosen subset of features; hence the name Random + Forest. Random forest is basically an advanced version of the bagging algorithm. In bagging, we build multiple decision trees with a bootstrapped training ...

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.

Read now

Unlock full access

More than 5,000 organizations count on O’Reilly

O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.

Julian F.

Head of Cybersecurity

I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.

Addison B.

Field Engineer

I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.

Amir M.

Data Platform Tech Lead

I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.

Mark W.

Embedded Software Engineer

The Human Factor in AI-Based Decision-Making

Publisher Resources

ISBN: 9781785884191

Cloud Computing

Data Engineering

Data Science

AI & ML

Programming Languages

Software Architecture

IT/Ops

Security

Design

Business

Soft Skills

Smarter Decisions - The Intersection of Internet of Things and Decision Science

by Jojo Moolayil

Ensemble modeling - random forest

What is random forest?

Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.