Bigger wins!

Sponsored By:
7:15-8:15am

Registration & Breakfast

8:15-8:30am

WELCOME REMARKS

[ Top of this page ] [ Agenda overview ]

8:30-9:30am

Keynote
Becoming an Ace with a Robot as your Wingman!

Humans and computers have strengths that are more complementary than alike – to the point where a sophisticated algorithm may be the best "2nd person" to put on a complex task. Yet, our and computer analytic weaknesses are surprisingly severe. To explore how to improve the man/machine partnership, we compare and contrast natural and artificial intelligence, with special attention to the growing realization of how challenging it is to think truly rationally.

9:30-10:10am

SAS Panel
Gaining Executive and Senior Management Support for Predictive Analytics

As government agencies face budget cuts, increased efficiency and effectiveness are in high demand. Predictive analytics can play a powerful role, but securing executive and senior management support can be difficult. Government organizations often recognize the importance of collecting and storing data, but it has limited value until enriched by advanced analytic efforts. Justifying the investment calls for a convincing demonstration of how predictive analytics can enhance performance and effectiveness. Our speakers will discuss how to successfully communicate the value of predictive analytics to mission-critical initiatives.

[ Top of this page ] [ Agenda overview ]

Sponsored By:
10:10-10:50am

BREAK

10:50am-12:00pm • Track1

Data Analytics to Improve Risk Management, Cyber Security, and DHS Policy

Hear from Department of Homeland Security executives about how they are using data analytics to focus on the biggest risks, to detect data breaches and intrusions, and to inform DHS risk management policies and programs. Topics explored will include using analytics to generate analytically defensible policy changes to manage risks, find the needles in a haystack with cyber security intrusions, generate predictive risk assessments, and to become more real-time in responding to incidents. Panelists will discuss how predictive analytics can be used in both the cyber and physical worlds and how to evolve risk assessment practices to include both retrospective and predictive elements.

10:50am-12:00pm • Track2

IRS Panel:
Strategic Workforce Analytics Tool at IRS

The IRS has developed a strategic workforce planning tool for the agency. We have modeled both workforce supply and workload demand and forecasted both through the next 5 to 10 years. The workforce supply models cover over 100,000 employees distributed among over 500 cities thoughout the country. The models forecast future employee attrition and movement within IRS business units, workstreams, and positions using statistical equations and probability matricies. The workload demand models forecast future tax administration demands on the Service across 19 major workstreams. We aligned the models using FTE multipliers to estimate future workload coverage scenarios.

[ Top of this page ] [ Agenda overview ]

Sponsored By:
12:00-1:15pm

Lunch

1:15-2:00pm

Keynote
Making Analytics Pay

Data Analytics remains a hot topic and is a priority discipline for both government and industry. Our ability to capture and create meaningful information from data never has been greater, yet the Holy Grail remains our ability to transform this data into meaningful and sustainable business result. Today's analytic organizations need to focus beyond the tools and methodologies of their trade to create closer and deeper relationships with their customers as well as embrace a strategic role in their agencies – and to structure themselves accordingly. Mr. Silverman will share the mission, structure and approach being pioneered at the IRS to make data analytics pay, and will relay core lessons learned about sustainably building analytic capabilities into the Service more broadly.

[ Top of this page ] [ Agenda overview ]

2:00-2:30pm

New Methods for Developing Limited Datasets for Predicting USMC Combat Losses – JMP, Booz Allen, and USMC

With the planned decreases to the overall Department of Defense budget, there is an increasing emphasis on achieving more efficient net asset management throughout all services. The recalculation of key variables used for the development of the USMC's War Reserve Materiel Requirement provides an opportunity to reduce total Marine Corps ground equipment liabilities while still ensuring the enterprise has the combat-essential equipment necessary to conduct operations. Predictive modeling of ground equipment combat losses observed in Operation Iraqi Freedom (OIF) and Operation Enduring Freedom (OEF) is explored to develop replacement rates or Combat Active Replacement Factors (CARF). The resulting interactive predictive modeling capability, called CARF Statistical Analysis Tool (CARF-STAT), is implemented in JMP Pro. With CARF-STAT, the Marine Corps is able to more accurately and consistently predict CARF values and effectively develop long-term plans for budget and equipment replacement.

2:30-3:00pm • Track1

Audit Analytics: Prioritizing Audits of Facility Lease Renewals

The US Postal Service actively manages over 25,000 leases and must renew about 4,000 of the leases each year. Most of the 4,000 leases have 24 or more months remaining on their lease term. Analytics is used to prioritize the leases for renewal review and assist in audit planning by highlighting areas or controls that are outside of normal operating conditions.

The resulting models can be continuously scored to highlight unexpected changes that might trigger an audit outside of the annual plan and the results can be used to support management initiatives and measure their effectiveness over time. Compared with typical fraud models that are focused on transactions or individual entities, audit models are focused on a collection of entities or behaviors of entities associated with operational controls.

2:30-3:0pm • Track2

Future Data Standardization Mandates and What They Mean for Analytics

The power of predictive analytics grows as more data sets are standardized using common formats and identifiers. Standardization reduces the time and effort needed to prepare data sets for the deployment of predictive analytics. Common formats such as XBRL, currently deployed by the SEC and the FDIC, and common identifiers such as the Legal Entity Identifier and OMB's planned federal award identifier promise to enhance the power of predictive analytics in the future. Congress is moving toward a comprehensive standardization mandate for federal spending data with the House passage and Senate consideration of the DATA Act. Congress is also expected to consider new standardization mandates for regulatory filings. To support such mandates - and ensure that they are sustainable and realistic - technology companies have formed the Data Transparency Coalition. This session will provide an overview of legislation the Coalition supports and describe how future mandates could create new opportunities for predictive analytics.

[ Top of this page ] [ Agenda overview ]

Sponsored By:
3:00-3:35pm

BREAK

3:35-4:05pm • Track1

Maximizing CMS Data for Internal and External Users

This session will provide an overview of the ways in which CMS is employing advanced analytics and data mining to target problem areas and develop interventions. The session will also discuss efforts CMS is undertaking to provide more and better data and information products to external users.

3:35-4:05pm • Track2

5 Myths of Predictive Analytics

Predictive Analytics is powerful, it can help you predict an event or a behavior at a an individual customer level. It can help you spot golden nuggets from the deep-wide-big data ocean; But is also one of the techniques which is not very well understood. With all the recent buzz about Predictive Analytics, it does seems like a new technique in the tool box. Is that so? In this keynote, we will ground ourselves in the reality of building and maintaining an impactful Predictive Model and explore questions like

1. Is Predictive Analytics new?
2. Is it a crystal ball?
3. Is it perfect?
4. Can it be built quickly and cheaply?
5. Is it going to solve all my business problems?
6. Does it always work?
7. Can anybody learn how to build a model?

[ Top of this page ] [ Agenda overview ]

4:05-5:15pm

Panel Discussion
How to Start a Data Analytics Group

5:15-6:15pm

NETWORKING RECEPTION

[ Top of this page ] [ Agenda overview ]

Sponsored By:
7:15-8:15am

Registration & Breakfast

8:15-9:00am

Keynote
Recovery Accountability and Transparency Board Use of
Analytics and Data for Fraud Prevention

The unique combination of advanced analytical tools, inquisitive analysts and a broad array of data sources has made the Recovery Board successful in preventing fraud under the $840 billion dollar Recovery Act program. The approaches and tools that are used by the Recovery Board will be discussed. The Board's FederalAccountability platform, and specific analytical tools available for analysts use on structured and unstructured data will be highlighted along with real world examples.

The presentation will also focus on data and data issues. The availability and quality of data are major issues for fraud analytics. Discussion topics will include data access, cooperation across organizations, data gaps and legal issues such as computer matching, privacy, and trade secrets. Approaches to decoupling data from systems and applications to improve availability and analysis will be covered.

[ Top of this page ] [ Agenda overview ]

9:00-9:45am

Healthcare Fraud - Detecting Workers' Compensation Fraud with Predictive Analytics

9:45-10:15am

Applying a Risk Intelligent Modeling (RIM) Approach to Financial Management Transformation

The United States Coast Guard (USCG) is engaged in a multi-year financial management transformation project to create world-class accounting, business information, and financial processes. In this session, you will learn how the USCG used a RIM approach to develop accounting estimates with sufficient monetary precision and rigor to be acceptable to financial accountants and auditors. Leveraging exploratory data analytics, sophisticated risk-based statistical sampling and regression techniques, the USCG was able to estimate account balances to work around inaccurate and inconsistent system records. Additionally, RIM was used to sustain remediation efforts and, drive process and internal control improvements.

[ Top of this page ] [ Agenda overview ]

Sponsored By:
10:15-10:50am

BREAK

10:50-11:20am • Track1

Detecting Fraud, Waste, and Abuse with Predictive Analytics

10:50-11:20am • Track2

Forecasting Housing Subsidy Need in an Uncertain Economy

The US Department of Housing and Urban Development (HUD) has the mission is to create strong, sustainable, inclusive communities and quality affordable homes for all. One of its program provides housing to individuals through contracts with renters of large multi-family units (e.g. apartment complexes) who agree to rent to qualified tenants under a long term contract. The rents for individual units are set by a contract that may run many years. HUD will subsidy these rents based on a tenants ability to pay. As the tenants income varies, so too does the subsidy need. Whenever department budgets are submitted to congress, or incremental funding in needed, HUD Multi-family budget analysts must forecast the total subsidy needed for more than 17,000 contracts that supply over 1 million units to more than 2.2 million tenants. Over the past year, HUD has integrated its contract, tenant, cost, and other data to create a budget forecasting tool that accurately forecasts that need. Articulated with a range of options and parameters the integrated Budget Forecasting Model (iBFM) is a critical tool in the ever-changing budget formulation process.

The forecasting models include time-series forecasting, regression models, and data mining-based predictive models based on macro-economic data combined with housing unit descriptions and tenant demographics.

[ Top of this page ] [ Agenda overview ]

11:25-11:55am • Track1

Reducing the Time and Cost of Predictive Analytics

11:25-11:55am • Track2

Preventing Catastrophic Incidents By Predicting Where They Are Most Likely To Occur And Why

What if you could predict where a serious incident is most likely to occur and why in plenty of time to prevent it from happening? That is exactly what we are trying to do at the Pacific Northwest National Laboratory (PNNL). We mined a diverse set of existing data sources focusing on those that could provide information about workgroups in three key areas: past operational experience, staff engagement, and exposure to severe hazards. We then used statistical analysis and predictive modeling tools to process all the data and identify the workgroups at highest risk for a significant operational event.

[ Top of this page ] [ Agenda overview ]

Sponsored By:

12:00-1:00pm

LUNCH

1:05-1:50pm

Keynote
Predictive Modeling to Find Medicare Fraud, Waste, and Abuse at CMS

Mr. Nelson will discuss the current and future efforts of the Center for Program Integrity's Data Analytics and Control Group to use predictive analytics to detect and prevent Medicare and Medicaid fraud, waste, and abuse. Topics will include methodologies used by the Data Analytics and Control Group to build predictive models and other sophisticated analytics to find high-risk provider behavior and payments using Medicare FFS claims, provider screening, and a variety of other data and the challenges associated with measuring prevention actions.

[ Top of this page ] [ Agenda overview ]

1:50-2:20pm

Big Data Analytics Best Practices of Real-World Government Agencies

How can agencies move from "sense and respond" to "predict and act"? When is the last time you made a decision based on a predictive insight? Hadoop is the new buzzword, but what is the path to implementation and what are the real opportunities? Using real-world government examples, this session will detail best practices for Advanced Analytics as well as the technology, services and education needed to better understand how to utilize all of your data sources. Learn how you can accelerate the pace of innovation with a holistic approach to the Big Data opportunity.

[ Top of this page ] [ Agenda overview ]

2:20pm-2:30pm

The Many Facets of 'Predictive' Analytics in Government

In this very brief overview we'll look at applications of predictive analytics in which 'prediction' really isn't the focus. From modeling of extremely rare events to the detection of insider threat activity to the study of runway incursion incidents we'll see that the term doesn't begin to describe the broad utility of the science we've come to label predictive analytics.

Sponsored By:
2:30pm-3:05pm

BREAK

3:05-3:45pm • Track1

Applying Predictive Analytics to Public Sector Applications

Not all government applications of Predictive Analytics fall nicely within the common solutions enjoyed by our private sector colleagues. Within the public sector there are many applications for Predictive Analytics but we are often faced with unique challenges. This session will share some unique and interesting government case studies where predictive analytics has been successfully applied to government applications including threat prediction and prevention, the use of non-traditional data sources and modeling rare events.

3:05-3:45pm • Track2

GOT BOTS? A Machine Learning Technique For Effective Detection Of BotNets

Botnets are one of the most sophisticated and popular types of cyber warfare today. They allow hackers to take control of hundreds of computers at a time, and turn them into \\\"zombies\\\" to spread viruses and perform other computer and network infrastructure-crippling activities. Using a Botnet case study, we describe implementation of a Quality Threshold Clustering (QTC) approach to distinguish between Botnet IRC traffic and real IRC traffic. Additionally, this session will: (a) compare results to a K Nearest Neighbor clustering approach; and (b) propose a computer network defense analytic strategy for Botnet detection and mitigation.

[ Top of this page ] [ Agenda overview ]

3:50pm-4:50pm

Panel Discussion
Predictive Analytics from the CTO Perspective

4:50pm-5:00pm

Closing Remarks

[ Top of this page ] [ Agenda overview ]