Monday February 15, 2010
Full-day Workshop
Room: California Parlor
Putting Predictive Analytics to Work
- Registration and breakfast starts at 8:00am
- Workshop starts at 9:00am
- Morning Coffee Break at 10:30am - 11:00am
- Lunch provided at 12:30 - 1:15pm
- Afternoon Coffee Break at 2:30pm - 3:00pm
- End of the Workshop: 4:30pm
Speaker: James Taylor, CEO, Decision Management Solutions
Full-day Workshop
Room: SAS Training Center, 120 Kearny St. (between Post & Sutter), Suite 3400
Hands-On Predictive Analytics
- Registration and breakfast starts at 8:00am
- Workshop starts at 9:00am
- Morning Coffee Break at 10:30am - 11:00am
- Lunch provided at 12:30 - 1:15pm
- Afternoon Coffee Break at 2:30pm - 3:00pm
- End of the Workshop: 4:30pm
Speaker: Dean Abbott, President, Abbott Analytics
Tuesday February 16, 2010
8:00am-9:00am
Room: Regency Foyer
Registration & Continental Breakfast
9:00am-9:50am
Room: Concert Ballroom
Keynote
Eureka! Seven Innovative Applications of Predictive Analytics
The final frontier isn't prediction - it's the prediction of new things. In this keynote, Dr. Eric Siegel presents seven of the latest and greatest innovative applications of predictive analytics that benefit organizations in new and creative ways. These examples represent the virtually boundless incoming wave of new "killer apps." After this keynote, each example is presented in greater detail during individual sessions of this Predictive Analytics World's 2-day agenda.
Applying predictive analytics in a new way requires only two particulars: 1) a new thing to predict (be it an event, incident or the lack thereof) and 2) a value proposition that comes of acting upon these predictions to drive decisions more effectively.
Given this world of possibilities, the range of predictive analytics applications is quickly expanding. Data-driven models predict new things such as the reliability of hardware and corporate processes alike, and drive all kinds of organizational decisions, for the likes of air traffic management, military operations, and startup investment strategy.
Taking a step beyond standard business and marketing applications such as response and churn modeling, we practitioners stand to learn a lot from such engineering, scientific and other applications of predictive analytics.
Speaker: Eric Siegel, Ph.D., Conference Chair
[ Top of this page ] [ Agenda overview ]
9:50am-10:10am
Room: Concert Ballroom
Platinum Sponsor Presentation
Beyond the Four Walls of your Enterprise: What your Customers Tell You
Gaining key insights from data to drive better decisions is no longer limited to information stored within the four walls of your organization. Enabled by Web 2.0 techniques, you can now capture key customer behaviors and conversations taking place outside of your enterprise so that you can gain a more complete view of your customers, and make better decisions, at key "moments of truths," to drive unique competitive advantage.
This presentation aims at answering questions such as:
- In this New Data World, how can organizations establish multi-channel conversations with their customers?
- In the loud ambient noise of Web 2.0 what are the essential pieces of data that will allow enterprises building one-to-one relationships with their customers?
- What are the essential points of time, moments of truth, when organizations can derive the most out of their customer relationships?
Speaker: Erick Brethenoux, VP Corporate Development, SPSS
[ Top of this page ] [ Agenda overview ]
Handbook of Statistical Analysis and Data Mining Applications
by Robert Nisbet, John Elder, Gary Miner
Meet the authors - Purchase a signed copy
Morning Breaks - StatSoft Booth
10:10am-10:30am
Room: Gold Ballroom
Break / Exhibits / Book Signing
Delegates may choose to attend either session at this time.
Session Moderators: Track 1 - Eric Siegel; Track 2 - James Taylor
10:30am-11:20amRoom: Concert Ballroom
Track 1: Thought Leader
Case Study: IBM
Analytic Journeys
Many companies are adopting analytics, with the most sophisticated increasingly pushing predictive analytics to the point of contact, the very tip of their organizations. Based on research conducted with IBM and IBM clients, this presentation will show how companies in a variety of industries have made progress on their analytic journeys. While each industry, each company, is different, this presentation will describe the common steps on the journey to pervasive, actionable, predictive analytics.
Speaker: James Taylor, CEO, Decision Management Solutions
[ Top of this page ] [ Agenda overview ]
10:30am-11:20am
Room: Rose Ballroom
Track 2: Lifetime Value
Case Study: Group RCI
Membership Value Modeling with Soft Revenues and Uncertain Cost
We discuss the development of a Membership Lifetime Value Model for Group RCI's North American vacation exchange business. In particular, we blend standard LTV concepts like attrition and future revenue forecasts with soft components like the value received from a member's vacation ownership for exchange. In addition, as Group RCI improves its web presence, we face the challenge of forecasting members' channel behavior and resulting cost profiles along the continuum from web only to mixed web/call center to pure call center activity. As member behavior changes rapidly we describe techniques and lessons learned in cost forecasting with limited information.
Speaker: Michael Towns, Director, Inventory & Consumer Analytics, Group RCI
[ Top of this page ] [ Agenda overview ]
11:20am-12:20pm
Room: Concert Ballroom
Multiple Case Studies: U.S. DoD, U.S. DHS, SSA
Text Mining: Lessons Learned
Text Mining is the "Wild West" of data mining and predictive analytics - the potential for gain is huge, the capability claims are often tall tales, and the "land rush" for leadership is very much a race.
In solving unstructured (text) analysis challenges, we found that principles from inductive modeling - learning relationships from labeled cases - has great power to enhance text mining. Dr. Elder will highlight key technical breakthroughs discovered while working on projects for leading government agencies, including:
- Prioritizing searches for the Dept. of Homeland Security
- Quick decisions for Social Security Admin. disability
- Document discovery for the Dept. of Defense
- Disease discovery for the Dept. of Homeland Security
- Risk profiling for the Dept. of Defense
Dr. Elder will summarize, from these (and commercial) deployment experiences, the factors essential to a successful text mining project.
Speaker: John Elder, Ph.D., Elder Research, Inc.
[ Top of this page ] [ Agenda overview ]
12:20pm-1:20pm
Room: Gold Ballroom
Birds of a Feather Lunch
Find your clan of like-minded colleagues with whom to dine and discuss, comparing your organization's stories and challenges.
Discussion topics:
- Response modeling
- Churn modeling
- Product recommendations
SAS Lunch Topics:
- Rapid Predictive Modeling
- Structured and Unstructured Data Analysis
- Sentiment and Social Media Analysis
1:20pm-2:10pm
Room: Concert Ballroom
Keynote
Predictive Power Part II: Advanced Analytics in the New Data Economy
Welcome to the New Data Economy, bringing new rules, reason and requirements to the commercial data miner. What's new? The relentless focus on value for the consumer, corresponding new revenue models, and ubiquitous data collection capabilities are here to stay, delivering a whole new world of opportunity and challenge for the practitioner of advanced analytics.
Mobile. The mobile device delivers new power in its omnipresent capacity to collect behavioral data - including device use, sound (voice quality), and movement (geo-location) - and, in turn, to influence behavior. iPhone apps such as RedLaser are "killer," putting the consumer first by letting the user scan barcodes and compare prices and alternatives on the go. Mobile application Boy Ahoy proves looks also can "kill," combining online dating with geolocation so you can "predictively score" with the guy standing right next to you.
Explicit data collection. No enterprise is more consumer-centric than the social network, which first and foremost incents users to explicitly volunteer personal information, and goes from there to facilitate interactive behavior - such as virtual gifting. Explicit data empowers analysis beyond the implicitly-conveyed consumer data businesses traditionally track. Boundaries between the private and the public have disappeared or been re-negotiated: At the key of Facebook's functionality is the distribution of "news" in mutually agreed "friendships."
Virtual gifting. Revenue is generated from "thin air" - and new data is collected - with a new economy of intention amongst consumers such as Facebook users, who pay to communicate attention or demonstrate affection by buying virtual flowers and other such goodies composed of nothing more than pixels on the screen of the beloved and all of their and the gifter's friends, thereby providing a quickly-escalating revenue model that competes with paid ads.
Virtual currency. Similarly, online gamers log on for free, but fork over real cash to buy weapons and armor. Such virtual currency provides an incentive for users to exhibit behavior that is good for the ecosystem.
Social data. Telecom - both mobile and otherwise - and social networks deliver a rich new data set that reveals what can be predicted from the behavior of one's friends (see also the "Social Data" track of conference sessions, just 20 minutes after this keynote address).
Cross-vendor data. Social networks and travel option aggregators track consumer trends across vendor - Priceline is privy to more consumers' behavior than any individual hotel chain or airline. These are the new consumer-centric data collectors.
In this keynote, learn from the former Chief Scientist of Amazon.com how to:
- Position the deployment of predictive models within the New Data Economy so that new opportunities are leveraged, new value propositions are pursued, and the escalating requirements of commercial deployment are met.
- Focus on the new data opportunities, not only the analytical methods, given the New Data delivered by consumer-centric business and technology, such as mobile, explicit, social and cross-vendor data.
Speaker: Andreas S. Weigend, Ph.D., Former Chief Scientist, Amazon.com
[ Top of this page ] [ Agenda overview ]
2:10pm-2:30pm
Room: Concert Ballroom
Platinum Sponsor Presentation
Predictive analytics and text - what the "experts" haven't told you
The advent of social media sites has pulled text analytics front and center. Effective predictive models using text depends on the quality of the data. This presentation will cover best practices as you embark on text analytic initiatives, covering approaches to data quality, categorization/classification, and sentiment on a variety of data sources. You'll find out what you need to know to successfully derive insight from text - not just once, but consistently over time, in a dynamic environment.
Speaker: Manya Mayes, Chief Text Analytics Strategist, SAS
[ Top of this page ] [ Agenda overview ]
2:30pm-2:40pm
Room: Gold Ballroom
Session Break
Delegates may choose to attend either session at this time.
2:40pm-3:30pm
Room: Concert Ballroom
Track 1: Social Data
Case Study: 1-800 Flowers
Leveraging Social Media Data to Manage Fraud Risk
With nearly two-thirds of the world's internet population posting content on social networks, blogs, photo sharing sites and more, social media is one of the richest pools of consumer data - and a substantial amount of this data is public, including demographics, interests, and more. This information can be leveraged to help companies identify fraud.
This session will explore how companies are now incorporating data such as friend counts, social network memberships, and other public data to score consumer risk and prevent fraudulent transactions online.
Speaker: Auren Hoffman, CEO, Rapleaf
[ Top of this page ] [ Agenda overview ]
2:40pm-3:30pm
Room: Rose Ballroom
Track 2: Text Analytics; Call Centers
Case Study: A Fortune 500 global technology company
Rules Rule: Inductive Business-Rule Discovery in Text Mining
Text mining remains at the leading edge (rather than the mainstream) of analytics for the corporate world, largely because of the complexities associated with how language is used. Words and phrases in a corporate lexicon can be used ambiguously, inconsistently, and incorrectly, making it difficult but not impossible for a human to understand. However, for predictive analytics, these ambiguities must be overcome so that algorithms can be applied consistently to historic data.
Call center data is no exception to these problems. A Fortune 500 global technology company applied text mining to their help desk calls related to the repair of supported devices. Complexities included the usual text ambiguities and spelling errors, but also included variable English terms and abbreviations as used in foreign countries. By incorporating a combination of manual text extraction by domain experts with automated machine learning using decision trees, an operational system of business rules was developed that exceeds specifications for profitable identification of parts needed for repairs.
Speaker: Dean Abbott, President, Abbott Analytics
[ Top of this page ] [ Agenda overview ]
Delegates may choose to attend either session at this time.
3:30pm-4:20pm
Room: Concert Ballroom
Track 1: Social Data
Case Study: A Leading North American Telecom
The Social Effect: Predicting Telecom Customer Churn with Call Data
A major North American telecom sought to better understand factors driving its customer churn. We trained a predictive model on several TB of in-house call history data, associating features of call quality with customer attrition. This model will be applied to either reduce the costs of existing retention campaigns, or achieve higher retention at the same cost.
Speaker: Michael Driscoll, Ph.D., Principal, Founder, Dataspora LLC
[ Top of this page ] [ Agenda overview ]
3:30pm-4:20pm
Room: Rose Ballroom
Track 2: Call Centers & Customer Service
Case Study: Canadian Automobile Association
Optimizing Customer Service levels through Predictive Analytics
CAA South Central Ontario provides emergency roadside services to its members. A survey to these members was conducted to determine overall satisfaction levels. The results of this survey were used to build predictive analytic solutions that would ultimately increase the overall level of member satisfaction. There are multiple stages of interaction between the customer service rep and the member with varying data constraints between each stage.Our challenge here was to build a unique set of predictive solutions that could be deployed at each one of these stages within the process.
Speaker: Richard Boire, Partner, BoireFillerGroup
[ Top of this page ] [ Agenda overview ]
4:20pm-4:50pm
Room: Gold Ballroom
Break / Exhibits
Delegates may choose to attend either session at this time.
4:50pm-5:40pm
Room: Concert Ballroom
Track 1: Innovative Applications
Case Study: Younoodle
The Mathematics of Innovation: Predicting Startup Success
In the world of venture capital, it is often said that it is nearly impossible to predict the future success of an early stage startup company. Yet this assumption has remained largely untested due to a lack of reliable data and insufficient analytic tools. This session will cover the work done by Younoodle on addressing the problem of predicting the success of startup companies. Using innovative methods we have built the world's largest database of detailed startup information covering over 60,000 companies and 350,000 people. Drawing on this rich stream of data we have built three predictive models that enable better decisions to be made in this space.
The first model uses key information about the founding team to produce a team score that predicts future success based on the individuals, their backgrounds and relationships. The second model uses a range of real-time metrics to predict the current valuation of high growth potential private companies. The final model utilizes network theory to map a startups competitive landscape. By uncovering mathematical patterns within our data and creating models to exploit these we have developed predictive models that enable better decisions to be made in this $100bn+ global industry.
Speaker: Sean Gourley, Ph.D., Co-founder and Director of Data-tools, Younoodle
[ Top of this page ] [ Agenda overview ]
4:50pm-5:40pm
Room: Rose Ballroom
Track 2: Call Centers & Customer Service
Case Study: BBC
Identifying and Helping the Most Vulnerable
The UK Switchover Help Scheme was set up by the BBC, through an agreement with the UK Government, to offer practical help at switchover to older and disabled people who may face greater barriers in switching to digital TV. This session will describe how we have worked with the scheme to predict (identify), and help, those UK citizens who, for a variety of reasons, are most in need of assistance.
Central to the modeling effort is a requirement to maintain a high level of data privacy to protect the anonymity of the most vulnerable members of our society
Speaker: John McConnell, Director Consultant, Analytical People
[ Top of this page ] [ Agenda overview ]
Delegates may choose to attend either session at this time.
5:40pm-6:30pm
Room: Concert Ballroom
Track 1: Innovative Applications
Case Study: Visa
Building 10,000 Predictive Models: Scaling Health & Status Models to Large, Complex Systems
As systems, such as payments systems, online ad servers and social network applications, grow larger and more complex, it becomes challenging to monitor their health and status. These types of systems can contain thousands of different data feeds, data flows and processes. A problem with just one of them can interrupt payments, ads, and status updates, respectively. Often there are hourly, daily, weekly and seasonal variations in the data that complicates the identification of potential problems. We describe a methodology using predictive models for quickly detecting potential problems. We illustrate the talk using a case study from Visa.
Speaker: Robert Grossman, Managing Partner, Open Data Group
Speaker: Joe Bugajski, Senior Analyst, Burton Group
[ Top of this page ] [ Agenda overview ]
5:40pm-6:30pm
Room: Rose Ballroom
Track 2: Market Mix Modeling
Case Study: Walmart Financial Services
Market Mix Modeling in Retail Financial Services
Marketing Mix (MMX) models have become commonplace in B2C industries to determine the impact of various marketing vehicles on sales and product movement. With the economic downturn, understanding the impact of Marketing and Events on the bottom line has never been more topical. Key questions are how to validate the accuracy of MMX models, how should MMX models be adapted for decision making, and how to gain executive buy-in for MMX as a tool in Marketing Finance. A version of MMX modeling has proven useful for Walmart Financial Services to calculate the effect of different promotional activities on sales.
Speaker: Randall Anzalone, Sr. Finance Manager, Walmart Financial Services
Speaker: Dhiraj Rajaram, CEO, Mu Sigma Inc.
[ Top of this page ] [ Agenda overview ]
6:30pm-7:30pm
Room: Gold Ballroom
Reception
7:30pm-10:00pm
Room: Concert Ballroom
useR Meeting
Featured speaker: John Chambers
R is an open source programming language for statistical computing, data analysis, and graphical visualization. R has an estimated one million users worldwide, and its user base is growing. While most commonly used within academia, in fields such as computational biology and applied statistics, it is gaining currency in commercial areas such as quantitative finance and business intelligence.
Among R's strengths as a language are its powerful built-in tools for inferential statistics, its compact modeling syntax, its data visualization capabilities, and its ease of connectivity with persistent data stores (from databases to flatfiles).
In addition, R is open source nature and extensible via add-on "packages" allowing it to keep up with the leading edge in academic research.
For all its strengths, though, R has an admittedly steep learning curve; the first steps towards learning and using R can be challenging.
Click here for more information about this useR Meeting
[ Top of this page ] [ Agenda overview ]
7:30pm-10:00pm
Room: Rose Ballroom
Bay Area SAS Users Group Meeting
How Can SAS Analytics Solve Your Business Problems? Let Me Show You!
For more than three decades, customers have used analytical software from SAS to solve complex business problems. Key developments in recent years include:
- More powerful algorithms to address scalability as well as performance
- More flexible models to handle increasingly complex data
- Advanced visualization capabilities suited to a variety of analyses and audiences
- The deployment of analytical algorithms across the enterprise
This presentation will use a variety of examples to illustrate the ways in which SAS analytical products can meet your evolving business needs.
Click here for more information about this SAS Users Group Meeting
Speaker: Tonya Etchison Balan, Manager, Analytics Product Management, SAS Institute
[ Top of this page ] [ Agenda overview ]
Wednesday February 17, 2010
8:00am-9:00am
Room: Regency Foyer
Registration & Continental Breakfast
9:00am-9:50am
Room: Concert Ballroom
Keynote
Response Modeling is the Wrong Modeling: Maximize Impact With Net Lift Modeling
The true effectiveness of a marketing campaign isn't response rate! It's the incremental impact - that is, additional revenue directly attributable to the campaign that would not otherwise have been generated. Yet traditional targeting criteria are often designed to find clients that are interested in the product, but would have bought it whether or not they received a promotion. In such cases, the incremental impact is insignificant and the marketing dollars could have been spent elsewhere.
Net Lift Models are designed to maximize incremental impact by targeting the undecided clients that can be motivated by marketing. These "swing customers" are akin to the swing states of a presidential election; data miners could learn a lot from presidential campaigns.
Beyond targeted marketing, Net Lift methodology delivers tremendous performance improvements for deployed churn models - retaining "savables" while avoiding the adverse "reverse" affects retention outreach triggers for some customers - as well as other innovative business applications of this advanced analytical method.
This keynote will demonstrate how to build Net Lift Models (also referred to as Uplift or Incremental Lift) that optimize the incremental impact of marketing campaigns, discussing the pros and cons of multiple core analytical approaches.
Speaker: Kim Larsen, Charles Schwab & Co
[ Top of this page ] [ Agenda overview ]
9:50am-10:00am
Room: Concert Ballroom
Gold Sponsor Presentation
Next Generation Insight
Netezza, the data warehouse appliance leader, has introduced an advanced analytics appliance that moves complex analytics next to the data to deliver 10-100x performance increases. This high performance, scalable advanced analytics appliance includes an extensible development environment to meet the diverse needs of any organization and a starter-kit with embedded analytics.
Netezza has partnered with SAS, the leader in analytics, to embed the SAS Scoring Accelerator into the Netezza appliance. The SAS Scoring Accelerator for Netezza allows customers to use all the data - up to multi-petabytes available in the data warehouse - to make faster decisions on critical business drivers.
Speaker: Phil Francisco, VP, Product Management & Product Marketing, Netezza
Speaker: Tonya Balan, Manager of Analytics Product Management, SAS
[ Top of this page ] [ Agenda overview ]
Handbook of Statistical Analysis and Data Mining Applications
by Robert Nisbet, John Elder, Gary Miner
Meet the authors - Purchase a signed copy
Morning Breaks - StatSoft Booth
10:00am-10:30am
Room: Gold Ballroom
Break / Exhibits / Book Signing
Delegates may choose to attend either session at this time.
10:30am-11:20am
Room: Concert Ballroom
Lab Session:
Track 1: Sponsored by SAS
Text Analytics Lab: Case Study
Successful application of text analytics to real world business issues
Describing classification, prediction, sentiment analysis, and search techniques and their uses across different industries, this lab will identify some key advantages of using text analytics to address core business challenges. Mayes & Fiona will also address best practices for different industries and explore in more detail the ins and outs of social media analysis. Review a cross-industry case study to explore how other companies have been successful with these approaches. We will also dive deep into a particular application that is on the newer frontiers of text analytics.
Speaker: Manya Mayes, Chief Text Analytics Strategist, SAS
Speaker: Fiona McNeill, Product Marketing Manager, SAS
[ Top of this page ] [ Agenda overview ]
10:30am-11:20am
Room: Rose Ballroom
Lab Session:
Track 2: Sponsored by Vortex DNA
Better Debt Pricing and Recovery
As a new source of predictive data VortexDNA provides enhanced predictability for a wide range of human behavior from online advertising to insurance, healthcare to credit. In the last 12 months, 15 percent of American adults, or nearly 34 million people, have been late making a credit card payment and 8 percent (18 million people) have missed a payment entirely. In this session we outline how VortexDNA predictive data helped RML Limited, New Zealand's largest purchaser of debt, model a six-fold improvement in dollar recovered with the potential to boost profits and lead to the better pricing of debt across a range of sectors.
Speaker: Branton Kenton-Dau, CEO, VortexDNA
Speaker: Paul Hilton, COO, Receivables Management Limited
[ Top of this page ] [ Agenda overview ]
Delegates may choose to attend either session at this time.
Session Moderators: Track 1 - Eric Siegel; Track 2 - David Katz
11:20am-12:10pmRoom: Concert Ballroom
Track 1: Thought Leader
In-database Vs. In-cloud Analytics: Implications for Deployment
Many of the major database, analytical database and database appliance vendors have implemented (or bought) broad data mining and predictive analytics capabilities in their engines. A similar set of developments are happening in the cloud. This will have a substantial impact on how PA applications are built and deployed.
For example, the steps for extracting and preparing data are compressed, streamlining the effort. Quantitative methods implemented as functions within SQL lessen the need to gain proficiency in proprietary languages.
These developments make it more likely that predictive models can be embedded in operational applications more seamlessly and some examples will be given
Speaker: Neil Raden, CEO, Hired Brains
[ Top of this page ] [ Agenda overview ]
11:20am-12:10pm
Room: Rose Ballroom
Track 2: Innovative Applications
Case Study: Continental Airlines and PASSUR Aerospace
Predicting the Health of Air Traffic Systems
How analytics are helping airlines predict airspace performance so they can better plan and manage air traffic at the most congested and weather-challenged airports in the United States.
Speaker: Tim Grovac, Director of Information Technology, PASSUR Aerospace
[ Top of this page ] [ Agenda overview ]
12:10pm-1:20pm
Room: Gold Ballroom
Birds of a Feather Lunch
Find your clan of like-minded colleagues with whom to dine and discuss, comparing your organization's stories and challenges.
- Project management and organizational process
- Data preparation
- Core predictive modeling methods
SAS Lunch Topics:
- In Database Analytics
- Time Series Data Mining
- Twitter Analysis
Delegates may choose to attend either session at this time.
1:20pm-2:10pm
Room: Concert Ballroom
Lab Session
Track 1: Sponsored by SPSS, an IBM Company
Publishing Profitably with Predictive Analytics
Special Predictive Analytics World lab session featuring Analytical People and SPSS, an IBM Company
Learn how publishing organizations are using predictive analytics to improve their bottom line. Whether you're in the publishing industry - or just want to become more profitable using your existing customer data - this lab is a must-attend event. The lab features a demo of IBM® SPSS® Modeler software - the premier data mining workbench - and how it can help you:
- Retain existing customers and acquire new ones
- Improve customer "intimacy" to achieve greater loyalty
- Create the most profitable relationships with your vendors or customers
Speaker: Tim Daciuk, Director, Worldwide Demo Resources, SPSS
Speaker: John McConnell, Director, Analytical People
[ Top of this page ] [ Agenda overview ]
1:20pm-2:10pm
Room: Rose Ballroom
Lab Session
Track 2: Sponsored by Salford Systems
Addressing Analytics Challenges in the Retail and Insurance Industries
This live demo will give you insight as to how Analytics Software is used to develop predictive models in the retail and insurance industries. Greg Makowski, a Data Mining Consultant in Wells Fargo's Consumer & Small Business Deposit Risk Department, Noe Tuason, Customer Research Manager for Insurance at the California State Automobile Association (CSAA), and Mikhail Golovnya, Consulting Project Leader at Salford Systems, will speak to you about business challenges involving:
- Modeling Automation in an Enterprise Software Application
- Retaining High Profit while Minimizing Risk of Flight, and
- Pricing Models Using Claims Frequency and Severity Data.
In addition, Mikhail Golovnya will draw on a wealth of real-world consulting experience to illustrate the use of the software so that you can begin to create comparable models using your own data.
Speaker: Mikhail Golovnya, Consulting Project Leader, Salford Systems
Speaker: Noe Tuason, Customer Research Mngr., Insurance California State Auto Assoc.
Speaker: Greg Makowski, Data Mining Consultant, Wells Fargo
[ Top of this page ] [ Agenda overview ]
Delegates may choose to attend either session at this time.
2:10pm-3:00pm
Room: Concert Ballroom
Track 1: Cross-Sell & Recommendations
Case Study: PayPal (eBay)
Next Best Product Model: How to optimize on probability as well as profitability
PayPal Product Marketing had a question: Who to sell what product at what incremental profit?
In the payment industry where products are somewhat cannibalizing in nature, it is important to optimize the up-sell path not only on adoption rate but also on incremental profit.
This presentation will walk you through a simple but unique way to optimizing across multiple dimensions with rigor.
Speaker: Piyanka Jain, Sr. Manager, Paypal
[ Top of this page ] [ Agenda overview ]
2:10pm-3:00pm
Room: Rose Ballroom
Track 2: Innovative Applications
Case Studies: U.S. Army, and a large commercial satellite company
Reliability Analytics: Predicting Equipment Failure
Predictive analytics can reveal how and when valuable equipment will fail, allowing businesses to anticipate, rather than react to, problems. Improving the discipline of reliability modeling with modern predictive methods better answers questions such as:
- How long can I rely on this piece of equipment?
- When should I start its replacement?
- Do last year's issues tell us anything about next year's?
Data Miner Matt Strampe will demonstrate - on projects involving satellites and tanks - how maintenance and environmental data can be harnessed to improve reliability forecasting for expensive assets, uncover hidden causes of failure, and provide predictions at the unit level. Better data and analytics allow models to be more robust, so they can make useful long-term predictions even when known failure cases are very rare.
Speaker: Matthew Strampe, Programmer / Analyst, Elder Research, Inc.
[ Top of this page ] [ Agenda overview ]
3:00pm-3:30pm
Room: Gold Ballroom
Break / Exhibits
Delegates may choose to attend either session at this time.
3:30pm-4:20pm
Room: Concert Ballroom
Track 1: Cross-Sell & Recommendations
Case Study: AT&T
Market Simulation in the Telecommunications Sector
This session will cover choice modeling approaches to predict customer demand for telecommunications voice and data services. Choice models predict the probability that a customer will purchase a product or service. Choice modeling delivers a market simulation tool to aid in business decision making and predict changes in brand market share for a variety of strategic and tactical changes to products or services.
We outline how to successfully create the market simulation tool, using real world examples, and a market simulation tool will be demonstrated. We will also discuss benefits and shortcomings of alternative statistical models (Hierarchical Bayes and Latent-Class).
Speaker: John Colias, Ph.D., Senior Vice President, Decision Analyst
Speaker: Basile P. Goungetas, Ph.D., Consumer Product Planning & Research, AT&T Inc.
[ Top of this page ] [ Agenda overview ]
3:30pm-4:20pm
Room: Rose Ballroom
Track 2: Analytical Methods
Case Study: Deutsche Postbank
Model Retrospection - Boosting Your Learning Curve
Sound model validation is crucial for successful application of predictive modelling in CRM. During standard model validation runs a lot of useful information is generated. The talk presents a structured approach to gain insight from these data. Model retrospection is an innovative technique to show, on which customer segments a model under- or overperformed. This yields insight on customer behaviour on the one hand and suggests detailed ways to improve the models on the other. It can be applied to any kind of model.
Speaker: Franz Hofner, Analyst, Dt. Postbank AG
[ Top of this page ] [ Agenda overview ]
Delegates may choose to attend either session at this time.
4:20pm-5:10pm
Room: Concert Ballroom
Track 1: Cross-Sell & Recommendations
Case Study: Sun Microsystems
Time-Purchasing Analytics and Next-Logical Predictions: Lessons from the High-Tech B2B World
Understanding the current and historic customer buying behavior, especially in a complex business environment, is key to an active customer relationship. To that end, we devised a set of tools that, in great detail, analyzes customer behavior, particularly from transactional sources, and presents actionable outcomes for marketing and sales. The focus will be on Time-Purchasing Analytics (TPA) as well as Next-Logical Prediction (NLP). Both of these components are a subset of a larger customer-analytic framework (KaufPac) at Sun.
We present a variety of B2B case studies (global, local, product-based) on the specific application of these two components portraying successes, failures, communication issues, and adoption hurdles. In addition, practical issues of implementation, execution, and changes in mindset and attitude are covered.
Speaker: Peter Jaumann, Customer Analytics and Predictive Analytics, Sun Microsystems
[ Top of this page ] [ Agenda overview ]
4:20pm-5:10pm
Room: Rose Ballroom
Track 2: Analytical Methods
Case Study: Google
Massively Parallel Learning of Tree Ensembles to Predict Ad Quality
Classification and regression tree learning on massive datasets is a common data mining and discovery task, yet many state of the art tree learning algorithms require training data to reside in memory on a single machine. While more scalable implementations of tree learning have been proposed, they typically require specialized parallel computing architectures. In contrast, the majority of Google's computing infrastructure is based on commodity hardware. In this talk, we describe PLANET: a scalable distributed framework for learning tree models over large datasets. PLANET defines tree learning as a series of distributed computations, and implements each one using the MapReduce model of distributed computation. We show how this framework supports scalable construction of classification and regression trees, as well as ensembles of such models. We discuss the benefits and challenges of using a MapReduce compute cluster for tree learning, and demonstrate the scalability of this approach by applying it to a real world learning task.
Speaker: Sugato Basu, Ph.D., Senior Research Scientist, Google
[ Top of this page ] [ Agenda overview ]
Thursday February 18, 2010
Full-day Workshop
Room: California Parlor
The Best and the Worst of Predictive Analytics:
Predictive Modeling Methods and Common Data Mining Mistakes
- Registration and breakfast starts at 8:00am
- Workshop starts at 9:00am
- First AM Break from 10:00 - 10:15
- Second AM Break from 11:15 - 11:30
- Lunch from 12:30 - 1:15pm
- First PM Break: 2:00 - 2:15
- Second PM Break: 3:15 - 3:30
- Workshops ends at 4:30
Speaker: John Elder, Ph.D., CEO and Founder, Elder Research, Inc.
[ Agenda overview ]
