Transcript Slide 1

CHAPTER 4
ANALYTICS, DECISION SUPPORT, AND
ARTIFICIAL INTELLIGENCE
Brainpower for Your Business
Opening Case: Online Learning
Notice the increase in online learning and the decrease in traditional enrollments.
What are the causes of this: Universities offering more online classes? More
alternatives to Universities?
What are the business implications of this? More space on campus…what can
be done with that if it becomes available? Should colleges move towards online
courses?
Phases of Decision Making
Intelligence (diagnostic)
 Design (create solutions)
 Choice (select solution)
 Implementation (apply
solution)

May cycle between these
phases.
Types of Decisions

Structured decision : data is available and variables
are known. Decision is based on known criteria.

Semi-Structured decision: In between. Some
uncertainty.

Nonstructured decision: not sure what data to
collect.Variables affecting decision are not known. Not
sure what the success criteria may be.
Satisficing:
Not optimal decision but allows
us to satisfy predetermined
Criteria.
E.g., Salary at least $50,000.
Types of Decisions
Recurring decision
 Nonrecurring (ad hoc) decision

Decision Support Systems
Helps you analyze, but you must know how to solve the problem,
and how to use the results of the analysis



Model management component
Data management component
User interface management component
Components of a DSS
Geographic Information Systems

Geographic information system (GIS)

Spatial information is any information in map form
Used to analyze information, generate business intelligence,
and make decisions. It’s easier to see information on a map.
Suppose you are going to start a new store selling electronic
accessories. Where would you locate it?

www.historypin.com
https://www.youtube.com/watch?v=NqUSfjTSLyo
https://www.youtube.com/watch?v=tnRJaHZH9lo
DATA-MINING TOOLS AND MODELS
◦ Databases and DBMSs: operational data
◦ Multidimensional analysis tools: summarized
data
◦ Digital dashboards: managers can get realtime info and drill down or roll up
◦ Statistical tools: regression, summarization,
association rules, clustering
◦ GISs: See information on a map
◦ Artificial intelligence: Genetic algorithms,
neural nets
Data-Mining: Predictive Analytics

Predictive analytics
◦ highly computational data-mining technology
that uses information and business
intelligence to build a predictive model for a
given business application

Insurance, retail, healthcare, travel,
financial services, CRM, SCM, credit
scoring, etc
Data-Mining: Predictive Analytics
Example

Prediction goal
◦ What customers are most likely to respond to a
social media campaign within 30 days by purchasing
at least 2 products in the advertised product line?

Prediction indicators
◦
◦
◦
◦
Frequency of purchases
Proximity of date of last purchase
Presence on Facebook and Twitter
Number of multiple-product purchases
Data-Mining: Text Analytics

Text analytics
◦ uses statistical, AI, and linguistic technologies
to convert unstructured textual information
into structured information

Gaylord Hotels uses text analytics to
make sense of customer satisfaction
surveys
Data-Mining: Endless Analytics
Web analytics – understanding and optimizing Web
page usage
◦ Search engine optimization (SEO) – improving
the visibility of Web site using tags and key terms
 HR analytics – analysis of human resource and
talent management data
 Marketing analytics – analysis of marketing-related
data to improve product placement, marketing mix,
etc

Data-Mining: Endless Analytics
CRM analytics – analysis of CRM data to improve
sales force automation, customer service, and
support
 Social media analytics – analysis of social media
data to better understand customer/organization
interaction dynamics
 Mobile analytics – analysis of data related to the
use of mobile devices to support mobile computing
and mobile e-commerce (m-commerce)

Artificial Intelligence

Artificial intelligence (AI)

Types of AI systems used in business
1.
2.
3.
4.

Expert systems
Neural networks
Genetic algorithms
Agent-based technologies
AI systems deliver the conclusion (rather
than helping you analyze the options)
Expert Systems
Expert (knowledge-based) system
 Uses if-then rules…lots of them.
 Used for

◦ Diagnostic problems (what’s wrong?)
◦ Prescriptive problems (what to do?) E.g., should we
give a loan to someone. Or should a credit card
customer be allowed to run a charge? If customer
older than 40 AND customer income > 100,000 and
previous payments all on time THEN allow transaction.
Expert System: Components
1.
Information Types
◦
◦
◦
2.
People
◦
◦
◦
3.
Problem facts
Domain expertise
“Why?” information
Domain expert: know the rules to apply
Knowledge engineer: Capture the rules
Knowledge worker: use the rules
IT Components
◦
◦
◦
◦
◦
Knowledge acquisition
Knowledge base
Inference engine
User interface
Explanation module
Expert System: Components
What Expert Systems Can and Can’t
Do
 An
expert system can
◦ Reduce errors
◦ Improve customer service
◦ Reduce cost
 An
expert system can’t
◦ Use common sense
◦ Automate all processes
Neural Networks and Fuzzy Logic
Neural network (NN)
(or artificial neural network (ANN))
 Learns through training data
 Finds patterns
 Different from expert systems since rules
do not have to be spelled out here…it
discovers its own rules in the data!

The Layers of a Neural Network
Neural Networks Can…
 Learn
and adjust to new circumstances on
their own
 Take part in massive parallel processing
 Function without complete information
 Cope with huge volumes of information
 Analyze nonlinear relationships
Fuzzy Logic
 Fuzzy
logic
◦ a mathematical method of handling imprecise or subjective
information
 Used
to make ambiguous information such as “short”
usable in computer systems
 Applications
◦ Google’s search engine
◦ Washing machines
◦ Antilock brakes
Genetic Algorithms

Genetic algorithm (GA)

Takes thousands or even millions of possible solutions,
combining and recombining them until it finds the an
optimal solution

Needs a fitness function and encoding of the possible
solutions in terms of chromosomes. It then attempts to
find the best chromosomes, after reproduction and
mutation.

E.g., What is the best inventory level for engines for the
Ford F150?

Work in environments where no model of how to find the
right solution exists
Agent-Based Technologies

Intelligent Agents

Multi-Agent Systems
Intelligent Agents

Software that acts on our behalf
◦
◦
◦
◦
Information agents or shopping/buyer agents
Monitoring-and-surveillance agents
User or personal agents
Data-mining agents
Multi-Agent Systems
Biomimicry
 Swarm (collective) Intelligence.
Ants find shortest route to food by laying
pheromone trails that decay automatically if they
are unfit or the food supply is gone. Based on
this, they find theshortest path to the food.
Similarly, find simple rules that can create big
desirable patterns.
