Lec 16

BUSS 215
Big Data Analytics(II)
Jiye BAEK
Department of MIS
Korea University Business School
ANNOUNCEMENT
Free Riders
- Peer evaluation score will be applied
BUSS215 © Jiye BAEK

BIG DATA (II)
BUSS215: Introduction to MIS
Fundamentals IT Trends and Applications
Online Platform
Digital Economy
Business Analytics
Big Data Analytics

E-Commerce
Emerging Technologies
BUSS215 © Jiye BAEK 3

OVERVIEW
Big Data Analytics – Machine Learning
Artificial Intelligence
AI describes the different ways a machine interacts with the world around it.
Machine learning
Machine learning is a category of research and algorithms focused on finding
patterns in data and using those patterns to make predictions
Types of Machine Learning

Supervised Learning
Unsupervised Learning
Big Data Applications

A.I. (2001)
Source: https://www.youtube.com/watch?v=_19pRsZRiz4

BIG DATA ANALYTICS
At the most basic level, machines equipped with AI imitate the human thought
process such as the ability to identify an apple and orange
Source: https://www.qualcomm.com/news/onq/2017/09/01/whats-difference-between-artificial-intelligence-machine-learning-and-deep
Computer Science (traditional)
vs. Machine Learning

Source: https://www.cpcstrategy.com/blog/2019/01/examples-of-artificial-intelligence/ BUSS215 © Jiye BAEK 12
Source: https://www.redwoodalgorithms.com/ra-insights/2017/8/28/the-artificial-intelligence-that-is-shaping-what-we-watch
Source:https://www.redwoodalgorithms.com/ra-insights/2017/8/28/the-artificial-intelligence-that-is-shaping-what-we-watch
Source:https://https://www.pinterest.com/pin/725783296177702797/ BUSS215 © Jiye BAEK 17
BIG DATA ANALYTICS
Big Data Analytics
Big Data + Big Data Analytics

Refers to the DATA Methods of using Big Data to generate insight
• Leveraging a computer’s ability to learn

1 Machine Learning/Deep Learning without being explicitly programmed to solve
business problems
• Understanding value drivers from the ever-

IoT (Internet of Things) &
2 Sensor Analytics
growing network of connected physical
objects and the communication between them
• Mining product reviews to estimate

Modeling Willingness-to-
3 Pay
willingness-to-pay for product features
• Understanding human speech as it is spoken

Natural Language
4 Processing
through application of computer science, AI,
and computational linguistics
• Using distributed computing and machine

5 Analyzing Data @ Scale learning tools to analyze hundreds of
gigabytes of data
• Mining social data in real time to understand

Creating a Streaming Consumer when and where consumers are making
6 Behavior Data Lake choices
Source: PwC (2017)

BIG DATA ANALYTICS
Big Data Analytics
Structured Unstructured
Signal Analysis Cluster Analysis

Regression Time Series Analysis
Traditional
Distinguish between Discover meaningful

Discover relationships Discover relationships
noise and meaningful groupings of data
between variables over time
information points
A/B/N Testing
Experiment to find the Classification Simulation Modeling Spatial Analysis
most effective variation Organize data points Experiment with a Extract geographic or
of a website, product, into known categories system virtually topological information
etc
Visualization
Complex Event Sentiment Analysis
Use visual Predictive Modeling
Processing Extract consumer
representations of data Use data to forecast or
Combine data sources reactions based on
to find and infer behavior
to recognize events social media behavior
Emerging
communicate info
Network Analysis Deep QA Natural Language

Optimization
Discover meaningful Find answers to Processing
Improve a process or
nodes and human questions Extract meaning from
function based on
relationships on using artificial human speech or
criteria
networks intelligence writing
Source: PwC (2017)

BIG DATA ANALYTICS
https://youtu.be/f_uwKZIAeM0
BIG DATA ANALYTICS
Machine Learning
Machine learning is an approach, or subset, of AI, with an emphasis on

“learning” rather than just computer programming.
Rather than being programmed to recognize a

cat or human face, they can be trained with
images from which to generalize and recognize
specific objects.

WHAT IS MACHINE LEARNING?

MACHINE LEARNING
Machine Learning
With ML, a system improves its pattern recognition capabilities by learning
from its own mistakes.
Ø Which customers should they target with a special
offer, prior to contract expiration?

Traditional Solutions
Offer incentives to every customer before contract expires
Contact each customer to probe propensity to terminate contract
Send offer to customers with high contract value

Machine Learning – Big Data Solution

MACHINE LEARNING – EXAMPLE



Color (nm) Alcohol % Beer or Wine?

610 5 Beer
599 13 Wine
693 14 Wine

Split the data

Choosing a model

Training

Training

Training

Training

Evaluation

Evaluation

Evaluation

Prediction

WHAT IS MACHINE LEARNING?

FINDING PATTERNS AND MAKE PREDICTIONS

TYPES OF MACHINE LEARNING
Supervised Learning

Supervised Learning
Supervised learning is where you have input variables (x) and an output
variable (Y) and you use an algorithm to learn the mapping function from
the input to the output
Y=f(X)
The algorithm iteratively makes predictions on the training data and is

corrected. Learning stops when the algorithm achieves an acceptable level of
performance.
In supervised learning, a model is able to predict with the help of labeled

dataset

Supervised Learning
Labeled data

Two types of supervised learning
• Classification: A classification problem is when the output variable is a

category, such as “red” or “blue” or “disease” and “no disease”.
• Regression: A regression problem is when the output variable is a real
value, such as “dollars” or “weight”.

• Classification

• Regression: e.g.) Suppose company employees have rated their

satisfaction on a scale of 1 to 100









CLASSIFICATION OR REGRESSION?
Spam Detection à
House Price Prediction à
Churn Prediction à
Stock Price Prediction à

Unsupervised learning where you only have input data (X) and no
corresponding output variables.
The goal for unsupervised learning is to model the underlying structure or

distribution in the data in order to learn more about the data.
• Algorithms are left to their own devises to discover and present the interesting
structure in the data.
In Unsupervised Learning, the algorithms is trained using data that is

unlabeled

Unlabeled data

Two types of unsupervised learning
• Clustering: A clustering problem is where you want to discover the inherent

groupings in the data, such as grouping customers by purchasing behavior.
• Association: An association rule learning problem is where you want to
discover rules that describe large portions of your data, such as people that
buy X also tend to buy Y.

• Clustering: e.g.) Suppose a telecom company wants to reduce its

customer churn rate by providing personalized call and data plans

• Association

BIG DATA ANALYTICS
https://www.youtube.com/watch?v=HcqpanDadyQ&list=PLIivdWyY5sqJxnwJhe3etaK7utrBiPBQ2
BIG DATA ANALYTICS
Machine Learning – Example

Create machine learning model to distinguish homes in New York from homes
in San Francisco
http://www.r2d3.us/visual-intro-to-machine-learning-part-1/

BIG DATA ANALYTICS
Big Data Analytics – Deep Learning
Deep Learning
A subset of ML, Deep Learning uses massive amounts of data and computing
power to simulate Deep Neural Networks.
• Neural networks imitate the human brain’s connectivity, classifying data sets and
finding correlations between them.
• With its new found knowledge, the machine can then apply its insights to other data
sets.
A machine can process large amounts of data, recognize complex patterns,
and provide more detailed insights.
BIG DATA ANALYTICS
Machine Learning vs. Deep Learning
Machine Learning
Starts with relevant features being manually extracted from images
The features are then used to create a model that categorizes the objects in
the image.
Deep Learning
Relevant features are automatically extracted from images
“end-to-end learning” – where a network is given raw data and a task to
perform, such as classification, and it learns how to do this automatically.

BIG DATA ANALYTICS
Machine Learning vs. Deep Learning
https://youtu.be/-SgkLEuhfbg
BIG DATA ANALYTICS APPLICATIONS
Big Data Analytics – Financial Services
Product Development
Analyze consumer habits and behaviors to improve insurance offerings
through better risk measures
Sales and Marketing

Evaluate market trends and consumer spending habits to improve loan and
credit card offers
Operations
Use demographics, consumer spending habits and social media data sets to
identify fraud from behavioral patterns
Service
Monitor customer sentiment and social interactions to customize the banking
experience and interaction
Big Data Analytics– Financial Services
https://www.youtube.com/watch?v=1RYKgj-QK4I
Big Data Analytics– Healthcare
Product Development
Analysis of disease trends and clinical studies to direct and optimize R&D
efforts for drug discovery
Sales and Marketing

Analysis of patient history and disease patterns to target patients with care or
lifestyle improvements
Operations
Utilize growing claims data and patient demographics to optimize treatment
selections
Service
Improve treatments and management of chronic illness through data collected
from patient monitoring
Big Data Analytics– Healthcare
https://youtu.be/wOwept5WlWM
Big Data Applications – Retail
Product Development
Use consumer purchase data, social networks and geo location data to
optimize product mix
Sales and Marketing

Utilized data generated from consumer mobile devices for location based
sales offerings
Operations
Evaluate store traffic patterns to improve product placement and point of sales
processes
Service
Evaluate online and real time sentiment to improve and tailor in store and post
sale experiences
Big Data Applications – Retail
https://www.youtube.com/watch?v=7wskKs85Yu8
Big Data Analytics in Practice
Business Need Big Data Analytics Impact
Greater tailoring of credit card Statistical model based on public Net revenue grew at a CAGR
offers to fit customer needs credit and demographic data to target of 32% from 1994 to 2003;
customized products to customers prompted competitors to shift
focus to data and analytics
Data-enabled engine prognostics, Analysis of sensor data from hundreds Over 70% annual revenue
monitoring, maintenance and of sensors in 4,000 engines to identify from the aircraft engine
repair and solve issues weeks in advance division attributable to this
service
Search-to-purchase conversion Semantic search, which enables Increases 10-15% the
by anticipating intent of a shopper’s discovery using algorithms that rank likelihood that a customer will
search and delivering relevant results via social signals from around complete their purchase –
results the web translating to millions of
dollars in revenue
Transformation from subscription Analysis of data from 66 million Revenue and subscriber
streaming service to original subscribers’ viewing habits and base increased by 15% and
content producer preferences 9% respectively in 2013
Leverage Internet of Things (IoT) Launched software to help airlines and Estimated 1% reduction in
by connecting machines to facilitate railroads move their data to the cloud fuel costs, projected to save
data-enabled prognostics, increase and predict mechanical malfunctions, the airline industry $30 billion
efficiency and reduce downtime improve safety, and reduce trip over 15 years
cancellations and cost
Source: PwC (2017)

WRAP UP
Machine learning
Machine learning is an approach, or subset, of AI, with an emphasis on
“learning” rather than just computer programming.
Types of Machine Learning

Supervised Learning
Big Data Applications

RESOURCES
• Bain (2013) “The value of Big Data: How analytics
differentiates winners”
(http://www.mckinsey.com/Insights/Business_Technology/Views_from_the_front_lines_of_the_data_analyti
cs_revolution?cid=other-eml-alt-mkq-mck-oth-1403)
• BCG (2014) “Seven Ways to Profit from Big Data as a

Business”
(https://www.bcgperspectives.com/content/articles/information_technology_strategy_digital_e
conomy_seven_ways_profit_big_data_business/)
• McKinsey (2014) “Views from the front lines of the data-

analytics revolution”
(http://www.mckinsey.com/Insights/Business_Technology/Views_from_the_front_lines_of_the_data_analyti
cs_revolution?cid=other-eml-alt-mkq-mck-oth-1403)
• TED (2013) “Philip Evans: How data will transform business”

(https://www.ted.com/talks/philip_evans_how_data_will_transform_business)
• Machine learning sandbox à playground.tensorflow.org

Lec 16

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Lec 16

Uploaded by

Copyright:

Available Formats

BUSS 215

Big Data Analytics(II)

BUSS215 © Jiye BAEK

BUSS215: Introduction to MIS

Fundamentals IT Trends and Applications

Big Data Analytics

BUSS215 © Jiye BAEK 3

Types of Machine Learning

Big Data Applications

BUSS215 © Jiye BAEK 4

BUSS215 © Jiye BAEK 5

BUSS215 © Jiye BAEK 7

Big Data + Big Data Analytics

• Leveraging a computer’s ability to learn

• Understanding value drivers from the ever-

• Mining product reviews to estimate

• Understanding human speech as it is spoken

• Using distributed computing and machine

• Mining social data in real time to understand

Source: PwC (2017)

BUSS215 © Jiye BAEK 18

Signal Analysis Cluster Analysis

Distinguish between Discover meaningful

Network Analysis Deep QA Natural Language

Source: PwC (2017)

BUSS215 © Jiye BAEK 19

Machine learning is an approach, or subset, of AI, with an emphasis on

Rather than being programmed to recognize a

BUSS215 © Jiye BAEK 21

BUSS215 © Jiye BAEK 22

BUSS215 © Jiye BAEK 24

Contact each customer to probe propensity to terminate contract

Send offer to customers with high contract value

BUSS215 © Jiye BAEK 25

BUSS215 © Jiye BAEK 26

BUSS215 © Jiye BAEK 27

BUSS215 © Jiye BAEK 28

BUSS215 © Jiye BAEK 29

Color (nm) Alcohol % Beer or Wine?

BUSS215 © Jiye BAEK 30

Split the data

BUSS215 © Jiye BAEK 31

BUSS215 © Jiye BAEK 32

BUSS215 © Jiye BAEK 33

BUSS215 © Jiye BAEK 34

BUSS215 © Jiye BAEK 35

BUSS215 © Jiye BAEK 36

BUSS215 © Jiye BAEK 37

BUSS215 © Jiye BAEK 38

BUSS215 © Jiye BAEK 39

BUSS215 © Jiye BAEK 40

BUSS215 © Jiye BAEK 41

BUSS215 © Jiye BAEK 42

BUSS215 © Jiye BAEK 43

The algorithm iteratively makes predictions on the training data and is

In supervised learning, a model is able to predict with the help of labeled

BUSS215 © Jiye BAEK 44

BUSS215 © Jiye BAEK 45

Two types of supervised learning

• Classification: A classification problem is when the output variable is a

BUSS215 © Jiye BAEK 46

Two types of supervised learning

BUSS215 © Jiye BAEK 47

Two types of supervised learning