Transcript Wednesday January 11 , 2006

ISM 250: Data Mining and Business
Analytics
Lecture 1
Ram Akella
TIM/UCSC
[email protected]
650-279-3078
Course Structure
• Business Analytics and context
• Data Mining
• Integration of two closely related topics via
projects
• Theory, practice, industry/university experts
• Lectures, lab, projects
• Additional topic: Starting a company, and
technology management
Course Philosophy
• Instructor provokes thought, stimulates, integrates
• Students work ahead and after class, reading to
prepare, work on labs and projects with Silicon
Valley firms
• Web has a great deal of information
• Instructors role is to clarify, deepen understanding,
help digest and integrate, and achieve new
insights, problem solving and research
Grading
• (May alter to weight project/term/research paper more
heavil, if of sufficiently high quality)
• Weekly Homework on fundamental topics, quizzes/final,
Comprehensive Course Project/term paper (including
presentation to class)
• Homework: 20%
• Quizzes and final: 30%
• Project/Term paper: 50% (Project Schedule is fast!)
• Presentation: 10%
Student Interest in Course
•
•
•
•
I have a BS in ..X.., MS in…Y..
I would like to learn….
I would like to be able do…
I would like to possibly do a startup……
Student preparation: Name
Strong
Linear Algebra:
Bases & transforms
Orthogonalization
SVD
Statistics
Hypothesis testing
Regression
ANOVA
Stochastics
Markov Chains
Queueing
Weak
Course schedule, Modules and TAs
• 1-2 weeks – One student takes care of
everything
• Labs by more experienced students
Business Analytics
Business Functions
• Start -> Concept -> Product -> R&D-> Engineering
• Verification/validation/promotion -> Marketing (includes
pricing)
• Selling -> Sales
• Making (manufacturing)/delivering (services) ->
Operations/Supply Chain (Customer-supplier networks for
complex products)
• Money to make it all work -> Finance?
• IS/IT………..
• HR
Issues
• Learning customer preferences: Conjoint
analysis
• Demand-supply match of
– Designers and products/projects
– Orders and capacity
– Uncertainty (queueing and delays)
• => Constrained optimization
Issues (continued)
• Product portfolios to maximize profits
– Given resources
– Acquire resources
– Goal: Speed to market (to achieve premium)
• In finance and engineering
• Marketing
– Now, in E-Business: Web page layout optimization to
maximize yield and revenue
– Pricing
– Product diffusion: Bass Model
Issues (continued)
• In product development, operations, finance
– Options to acquire/buy/sell capacity, given uncertain
demand
• Tool kit: Stochastic Dynamic Programming (SDP)
and Real Options (Decision Trees)
• Use of SDP in Supply Chain Management
• Use of SDP and constrained optimization in
waterfall and spiral product development models
• Integration with data mining
Data Mining
•
•
•
•
•
•
•
Trends in demand
Changes
Anomalies
Quality characteristics: good/bad - classification
Price changes and clusters
Volume changes and clusters
Associations
Text Mining and Search
• Search for Product Component or Demand
– Right match by
•
•
•
•
Description (text)
Price
Quality
Volume
Data Mining (after next four slides)
Technology Ventures???
• Discuss after DM slides and lecture is
completed
Next Class: Reading R1
• BA: Conjoint Analysis
– Preferences in marketing
• DM: Metrics and data in data mining
• Products
– Manufacturing
– Knowledge
Next Class : Assignment 1
• Read “The Search” and summarize 10 key
ideas, rank ordered in descending priority
• Bullet point format is OK
Project
• FitMe: Presentation today at 7 pm