Three lessons learned from building a production machine learning system. Michael Manapat

Size: px
Start display at page:

Download "Three lessons learned from building a production machine learning system. Michael Manapat"

Transcription

1 Three lessons learned from building a production machine learning system Michael Manapat

2 Fraud Card numbers are stolen by hacking, malware, etc. Dumps are sold in carding forums Fraudsters use numbers in dumps to buy goods, which they then resell Cardholders dispute transactions Merchant ends up bearing cost of fraud

3 We train binary classifiers to predict fraud We use open source tools Scalding/Summingbird for feature generation scikit-learn for model training (eventually: github.com/stripe/brushfire)

4 1 Don t treat models as black boxes

5 Early ML at Stripe Focused on training with more and more data and adding more and more features Didn t think much about ML algorithms (tuning hyperparameters, e.g.) The deeper reasons behind any particular set of results Substantial reduction in fraud rate

6 Product development From a product standpoint: We were blocking high risk charges and surfacing just the decision We wanted to provide Stripe users insight into our actions reasons for scores

7 Score reasons True X < 10 False Y < 5 X < (20) 0.3 (30) 0.5 (10) 0.9 (40) X = 5, Y = 3: score = 0.1 Which feature is driving the score more?

8 Score reasons X <? Y < 5 X <? 0.1 (20) 0.3 (30) 0.5 (10) 0.9 (40) X =?, Y = 3: (20/70) * (10/70) * (40/70) * 0.9 = 0.61 Score Δ = holdout - original = = 0.51 Now producing richer reasons with multiple predicates

9 Model introspection If a model didn t look good in validation, it wasn t clear what to do (besides trying more features/data) What if we used our score reasons to debug model issues?

10 Take all false positives (in validation data or in production) and group by generated reason Were a substantial fraction of the false positives driven by a few features? Did all the comparisons in the explanation predicates make sense? (Were they comparisons a human might make for fraud?) Our models were overfit!

11 Actioning insights Hyperparameter optimization Recall Precision Feature selection

12 Summary Don t treat models as black boxes Thinking about the learning process (vs. just features and data) can yield significant payoffs Tooling for introspection can accelerate model development/ debugging Julia Evans, Alyssa Frazee, Erik Osheim, Sam Ritchie, Jocelyn Ross, Tom Switzer

13 2 Have a plan for counterfactual evaluation

14 December 31st, 2013 Train a binary classifier for disputes on data from Jan 1st to Sep 30th Validate on data from Oct 1st to Oct 31st (need to wait ~60 days for labels) Based on validation data, pick a policy for actioning scores: block if score > 50

15 Questions (1) Business complains about high false positive rate: what would happen if we changed the policy to "block if score > 70"? What are the production precision and recall of the model?

16 December 31st, We repeat the exercise from a year earlier Train a model on data from Jan 1st to Sep 30th Validate on data from Oct 1st to Oct 31st (need to wait ~60 days for labels) Validation results look ~ok (but not great) We put the model into production and the results are terrible

17 Questions (2) Why did the validation results for the new model look so much worse? How do we know if the retrained model really is better than the original model?

18 Counterfactual evaluation Our model changes reality (the world is different because of its existence) We can answer some questions (around model comparisons) with A/B tests For all these questions, we want an approximation of the charge/outcome distribution that would exist if there were no model

19 One approach Probabilistically reverse a small fraction of our block decisions The higher the score, the lower probability we let the charge through Weight samples by 1 / P(allow) Get information on the area we want to improve on

20 ID Score p(allow) Original Action Selected Action Outcome Allow Allow OK Allow Allow Fraud Block Block Block Allow Fraud Block Block Block Allow OK

21 ID Score P(Allow) Weight Original Action Selected Action Outcome Allow Allow OK Allow Allow Fraud Block Allow Fraud Block Allow OK Evaluating the "block if score > 50" policy Precision = 5 / 9 = 0.56 Recall = 5 / 6 = 0.83

22 The propensity function controls the exploration/ exploitation tradeoff Precision, recall, etc. are estimators Variance of the estimators decreases the more we allow through Bootstrap to get error bars (pick rows from the table uniformly at random with replacement) Li, Chen, Kleban, Gupta: "Counterfactual Estimation and Optimization of Click Metrics for Search Engines"

23 Summary Have a plan for counterfactual evaluation before you productionize your first model You can back yourself into a corner (with no data to retrain on) if you address this later You should be monitoring the production performance of your model anyway (cf. next lesson) Alyssa Frazee, Julia Evans, Roban Kramer, Ryan Wang

24 3 Invest in production monitoring for your models

25 Production vs. data stack Ruby/Mongo vs. Scala/Hadoop/Thrift Some issues Divergence between production and training definitions Upstream changes to library code in production feature generation can change feature definitions True vs. True

26 Logged scoring requests Domain-specific scoring service (business logic) Aggregation jobs Pure model evaluation service Aggregation jobs keep track of Overall action rate and rate per Stripe user Score distributions Feature distributions (% null, p50/p90 for numerical values, etc.)

27 Domain-specific scoring service (business logic) Pure model evaluation service Logged scoring requests Aggregation jobs (get all aggregates per model)

28 Summary Monitor the production inputs to and outputs of your models Have dashboards that can be watched on deploys and alerting for significant anomalies Bake the monitoring into generic ML infrastructure (so that each ML application isn t redoing this) Steve Mardenfeld, Tom Switzer

29 Don t treat models as black boxes Have a plan for counterfactual evaluation before productionizing your first model Build production monitoring for action rates, score distributions, and feature distributions (and bake into ML infra)

30 Thanks Stripe is hiring data scientists, engineers, and engineering

Counterfactual evaluation of machine learning models. Michael

Counterfactual evaluation of machine learning models. Michael Counterfactual evaluation of machine learning models Michael Manapat @mlmanapat ONLINE FRAUD ONLINE FRAUD ONLINE FRAUD RADAR Rules engine ( block if amount > $100 and more than 3 charges from IP in past

More information

Startup Machine Learning: Bootstrapping a fraud detection system. Michael Manapat

Startup Machine Learning: Bootstrapping a fraud detection system. Michael Manapat Startup Machine Learning: Bootstrapping a fraud detection system Michael Manapat Stripe @mlmanapat About me: Engineering Manager of the Machine Learning Products Team at Stripe About Stripe: Payments infrastructure

More information

Model Selection, Evaluation, Diagnosis

Model Selection, Evaluation, Diagnosis Model Selection, Evaluation, Diagnosis INFO-4604, Applied Machine Learning University of Colorado Boulder October 31 November 2, 2017 Prof. Michael Paul Today How do you estimate how well your classifier

More information

State of Card Fraud: 2017

State of Card Fraud: 2017 State of Card Fraud: 2017 What you need to know about the State of Fraud in 2017, including EMV updates, ongoing threats and how it s impacting financial institutions. EMV Adoption Update State of Card

More information

TSYS FORESIGHT SCORE SM

TSYS FORESIGHT SCORE SM TSYS FORESIGHT SCORE SM WITH FEATURESPACE Limit fraud losses. Generate more revenue. And give your customers the experience they want. Enhanced precision in detecting and preventing fraud Fine-tuned, real-time

More information

Predicting and Explaining Price-Spikes in Real-Time Electricity Markets

Predicting and Explaining Price-Spikes in Real-Time Electricity Markets Predicting and Explaining Price-Spikes in Real-Time Electricity Markets Christian Brown #1, Gregory Von Wald #2 # Energy Resources Engineering Department, Stanford University 367 Panama St, Stanford, CA

More information

Progress Report: Predicting Which Recommended Content Users Click Stanley Jacob, Lingjie Kong

Progress Report: Predicting Which Recommended Content Users Click Stanley Jacob, Lingjie Kong Progress Report: Predicting Which Recommended Content Users Click Stanley Jacob, Lingjie Kong Machine learning models can be used to predict which recommended content users will click on a given website.

More information

Deep Learning Hyperparameter Optimization with Competing Objectives

Deep Learning Hyperparameter Optimization with Competing Objectives Deep Learning Hyperparameter Optimization with Competing Objectives GTC 2018 - S8136 Scott Clark scott@sigopt.com OUTLINE 1. Why is Tuning Models Hard? 2. Common Tuning Methods 3. Deep Learning Example

More information

Lecture 6: Decision Tree, Random Forest, and Boosting

Lecture 6: Decision Tree, Random Forest, and Boosting Lecture 6: Decision Tree, Random Forest, and Boosting Tuo Zhao Schools of ISyE and CSE, Georgia Tech CS764/ISYE/CSE 674: Machine Learning/Computational Data Analysis Tree? Tuo Zhao Lecture 6: Decision

More information

Building A Modern Security Engineering Team Zane

Building A Modern Security Engineering Team Zane Building A Modern Security Engineering Team Zane Lackey zane@signalsciences.com @zanelackey April 2016 WHO IS THIS GUY ANYWAY? 1. Built and led the Etsy Security Team Spoiler alert: what this presentation

More information

Engineering Your Startup to Innovate at Scale. Randy linkedin.com/in/randyshoup

Engineering Your Startup to Innovate at Scale. Randy linkedin.com/in/randyshoup Engineering Your Startup to Innovate at Scale Randy Shoup @randyshoup linkedin.com/in/randyshoup Background VP Engineering at Stitch Fix o Combining Art and Science to revolutionize apparel retail Consulting

More information

Continuous Live Monitoring of Machine Learning Models with Delayed Label Feedback

Continuous Live Monitoring of Machine Learning Models with Delayed Label Feedback Continuous Live Monitoring of Machine Learning Models with Delayed Label Feedback Patrick Baier Lorand Dali Zalando Payments June 2018 OUTLINE Who we are and what we do Why we should monitor Prediction

More information

HOW TO USE AI IN BUSINESS

HOW TO USE AI IN BUSINESS HOW TO USE AI IN BUSINESS How to use AI in business 3 FOREWORDS 20 6 EXAMPLES OF USE CASES 4 INTRODUCTION TO AI 33 SUMMARY 10 MACHINE LEARNING IN PRACTICE 35 USEFUL LINKS 14 BEST PRACTISES FOREWORDS Artificial

More information

How to build and deploy machine learning projects

How to build and deploy machine learning projects How to build and deploy machine learning projects Litan Ilany, Advanced Analytics litan.ilany@intel.com Agenda Introduction Machine Learning: Exploration vs Solution CRISP-DM Flow considerations Other

More information

CALCULATING THE ROI OF LEAD NURTURING

CALCULATING THE ROI OF LEAD NURTURING Definitive Guide to Lead Nurturing Lead Calculating the ROI of Lead Nurturing In Part One of the Definitive Guide to Lead Nurturing, we learned the value of lead nurturing done well, and in Parts Two and

More information

Purchase Order, Requisitions, Inventory Hands On. Workshop: Purchase Order, Requisitions, Inventory Hands On

Purchase Order, Requisitions, Inventory Hands On. Workshop: Purchase Order, Requisitions, Inventory Hands On Workshop: Purchase Order, Requisitions, Inventory Hands In this follow up session to the Operations Changes in Purchase Order, Requisition, and Inventory Theory course, this hands on session will look

More information

It s been a long time coming but finally our Multi-Currency functionality is here!

It s been a long time coming but finally our Multi-Currency functionality is here! Sage One Multi-Currency Getting Started Guide It s been a long time coming but finally our Multi-Currency functionality is here! The functionality allows you to: Run your business in your country s main

More information

Leveraging Big Data For Payment Risk Management

Leveraging Big Data For Payment Risk Management Payments for platform businesses. New York 2014 Leveraging Big Data For Payment Risk Management John Canfield, VP Risk Management, WePay @JCRisk" June 11, 2014 Outline 1. Payments opportunity for the bottoms-up

More information

What s new in Machine Learning across the Splunk Portfolio

What s new in Machine Learning across the Splunk Portfolio What s new in Machine Learning across the Splunk Portfolio Manish Sainani Director, Product Management Bob Pratt Sr. Director, Product Management September 2017 Forward-Looking Statements During the course

More information

Modern Payment Fraud Prevention at Big Data Scale

Modern Payment Fraud Prevention at Big Data Scale This whitepaper discusses Feedzai s machine learning and behavioral profiling capabilities for payment fraud prevention. These capabilities allow modern fraud systems to move from broad segment-based scoring

More information

US state auditor deploys machine learning to tackle healthcare fraud at scale

US state auditor deploys machine learning to tackle healthcare fraud at scale US state auditor deploys machine learning to tackle healthcare fraud at scale This e-book focuses on how a US state auditor leveraged machine learning to detect and prevent fraud in healthcare. The biggest

More information

Chargeback Best Practices. September 7, 2016

Chargeback Best Practices. September 7, 2016 1 Chargeback Best Practices September 7, 2016 22 U.S. Payments Forum the cross-industry body focused on supporting the introduction and implementation of EMV and other new and emerging technologies that

More information

Using Data Analytics to Detect Fraud

Using Data Analytics to Detect Fraud Using Data Analytics to Detect Fraud Data Analysis Tests for Detecting Payroll and Expense Reimbursement Schemes 2018 Association of Certified Fraud Examiners, Inc. Payroll Schemes Fraudsters cause their

More information

Automating Operational Decisions in Real-time. Chris Sanden Senior Analytics Engineer

Automating Operational Decisions in Real-time. Chris Sanden Senior Analytics Engineer Automating Operational Decisions in Real-time Chris Sanden Senior Analytics Engineer About Me. Senior Analytics Engineer at Netflix Working on Real-time Analytics Part of the Insight Engineering Team Tenure

More information

3 Ways to Improve Your Targeted Marketing with Analytics

3 Ways to Improve Your Targeted Marketing with Analytics 3 Ways to Improve Your Targeted Marketing with Analytics Introduction Targeted marketing is a simple concept, but a key element in a marketing strategy. The goal is to identify the potential customers

More information

ADVANCED DATA ANALYTICS

ADVANCED DATA ANALYTICS ADVANCED DATA ANALYTICS MBB essay by Marcel Suszka 17 AUGUSTUS 2018 PROJECTSONE De Corridor 12L 3621 ZB Breukelen MBB Essay Advanced Data Analytics Outline This essay is about a statistical research for

More information

Machine Learning; forecasting safety risk and performance

Machine Learning; forecasting safety risk and performance Machine Learning; forecasting safety risk and performance Dr Claire Le Cras NATS Analytics Machine learning to predict safety events Risk Analysis Tool (RAT) is a methodology used to classify safety related

More information

Not All Machine Learning Systems Are Created Equal

Not All Machine Learning Systems Are Created Equal Not All Machine Learning Systems Are Created Equal Introduction The rise of e-commerce, alternative payment methods, and card not present transactions has caused the number of fraud losses by retailers

More information

Putting Card Fraud to the Fire. Diana Kern, AAP senior trainer

Putting Card Fraud to the Fire. Diana Kern, AAP senior trainer Putting Card Fraud to the Fire Diana Kern, AAP senior trainer Disclaimer: The following does not constitute legal advice. The information provided herein may not be applicable in all situations, should

More information

Online Credit Card Fraudulent Detection Using Data Mining

Online Credit Card Fraudulent Detection Using Data Mining S.Aravindh 1, S.Venkatesan 2 and A.Kumaravel 3 1&2 Department of Computer Science and Engineering, Gojan School of Business and Technology, Chennai - 600 052, Tamil Nadu, India 23 Department of Computer

More information

Rules vs Machine Learning

Rules vs Machine Learning For Sale Rules vs Machine Learning The future of fraud fighting is here. The Sift Science Trust Platform is the only fraud prevention platform leveraging Live Machine Learning, allowing you to reduce fraud,

More information

A stochastic approach to multi channel attribution in online advertising

A stochastic approach to multi channel attribution in online advertising A stochastic approach to multi channel attribution in online advertising Authors Venkatesh Sreedharan (venkateshw@freshworks.com Swaminathan Padmanabhan (swami@freshworks.com Machine Learning @Freshworks

More information

How to improve your AML detection? Christopher Ghenne Principal Manager Fraud & Security Intelligence EMEA

How to improve your AML detection? Christopher Ghenne Principal Manager Fraud & Security Intelligence EMEA How to improve your AML detection? Christopher Ghenne Principal Manager Fraud & Security Intelligence EMEA Years of 14,010 SAS employees worldwide 93 of the top 100 on the 40 #1 BUSINESS ANALYTICS companies

More information

Banking in the Balance: Security vs. Convenience. IBM Trusteer s Valerie Bradford on How to Assess Digital Identities

Banking in the Balance: Security vs. Convenience. IBM Trusteer s Valerie Bradford on How to Assess Digital Identities Banking in the Balance: Security vs. Convenience IBM Trusteer s Valerie Bradford on How to Assess Digital Identities In an interview about overcoming these challenges, Bradford discusses: The fundamental

More information

Using KPIs and Payments Data to Drive Your Business: A Fresh Perspective

Using KPIs and Payments Data to Drive Your Business: A Fresh Perspective Using KPIs and Payments Data to Drive Your Business: A Fresh Perspective Cybersource Payment Management Summit June 1, 2017 About Glenbrook Partners WHO ARE WE? Founded in 2001 and headquartered in Silicon

More information

Common Customer Use Cases in FSI

Common Customer Use Cases in FSI Common Customer Use Cases in FSI 1 Marketing Optimization 2014 2014 MapR MapR Technologies Technologies 2 Fortune 100 Financial Services Company 104M CARD MEMBERS 3 Financial Services: Recommendation Engine

More information

EMV in the U.S. Liability shift; what does this mean for the U.S.?

EMV in the U.S. Liability shift; what does this mean for the U.S.? EMV in the U.S. Liability shift; what does this mean for the U.S.? Questions and answers What the liability shift really means with regards to costs, risks and benefits. Fraud is on the rise in the U.S.

More information

Module 3: How to measure and maximize your Investment in the Addressable Platforms. Peter Vandre

Module 3: How to measure and maximize your Investment in the Addressable Platforms. Peter Vandre Module 3: How to measure and maximize your Investment in the Addressable Platforms Peter Vandre 1 Agenda Platform tracking innovation and impact on measurement Our integrated measurement framework How

More information

An IT Managers Guidebook to Implementing ITIL

An IT Managers Guidebook to Implementing ITIL 4 Steps to Awesome Service An IT Managers Guidebook to Implementing ITIL Defining strategy and process for continuous improvement An Entry Software Corporation Publication About the Author John McDonald

More information

Your Trusted Payment Gateway

Your Trusted Payment Gateway Your Trusted Payment Gateway Online financial transactions made more secured with this simplified payment gateway Latest encryption technology in RapidoPayments, tested against various security compromises

More information

Operations & Technology OPERATIONS & TECHNOLOGY CONFERENCE & EXHIBITION SEPTEMBER 11-13, 2017 THE WESTIN CHARLOTTE CHARLOTTE, NC

Operations & Technology OPERATIONS & TECHNOLOGY CONFERENCE & EXHIBITION SEPTEMBER 11-13, 2017 THE WESTIN CHARLOTTE CHARLOTTE, NC Operations & Technology OPERATIONS & TECHNOLOGY CONFERENCE & EXHIBITION SEPTEMBER 11-13, 2017 THE WESTIN CHARLOTTE CHARLOTTE, NC Conference and Exhibition The Future is Here: Do You Have the Tools Necessary?

More information

SellWise User Group. Thursday, July 16, Presenters. Will Atkinson, President CAP/Sellwise Mike Watkins, Member Care & Shared Services

SellWise User Group. Thursday, July 16, Presenters. Will Atkinson, President CAP/Sellwise Mike Watkins, Member Care & Shared Services SellWise User Group Thursday, July 16, 2015 Presenters Will Atkinson, President CAP/Sellwise Mike Watkins, Member Care & Shared Services SellWise User Group Thursday, July 16, 2015 Presenters Will Atkinson,

More information

Personal Financial Management

Personal Financial Management Personal Financial Management Note - Click the heading in each section to view the answers for each section. Personal Financial Management FAQs: 1. What is Personal Financial Management? 2. What are the

More information

Adapting software project estimation to the reality of changing development technologies

Adapting software project estimation to the reality of changing development technologies Adapting software project estimation to the reality of changing development technologies Introduction Estimating software projects where significant amounts of new technology are being used is a difficult

More information

Is Your Company Ready for Digital Transformation? Introducing the Digital Readiness Index

Is Your Company Ready for Digital Transformation? Introducing the Digital Readiness Index Is Your Company Ready for Digital Transformation? Introducing the Digital Readiness Index Sight Machine regularly meets with large manufacturers who are embarking on journeys of digital transformation.

More information

IBM. Processor Migration Capacity Analysis in a Production Environment. Kathy Walsh. White Paper January 19, Version 1.2

IBM. Processor Migration Capacity Analysis in a Production Environment. Kathy Walsh. White Paper January 19, Version 1.2 IBM Processor Migration Capacity Analysis in a Production Environment Kathy Walsh White Paper January 19, 2015 Version 1.2 IBM Corporation, 2015 Introduction Installations migrating to new processors often

More information

Reducing E-Commerce Fraud Loss by 25% The Role of Machine Learning, Artificial Intelligence and Real-Time Behavioral Analytics in Fighting Fraud

Reducing E-Commerce Fraud Loss by 25% The Role of Machine Learning, Artificial Intelligence and Real-Time Behavioral Analytics in Fighting Fraud Reducing E-Commerce Fraud Loss by 25% The Role of Machine Learning, Artificial Intelligence and Real-Time Behavioral Analytics in Fighting Fraud The Card-Not-Present Crisis Card-not-present (CNP) fraud

More information

Getting the Most out of Statistical Forecasting!

Getting the Most out of Statistical Forecasting! Getting the Most out of Statistical Forecasting! Author: Ryan Rickard, Senior Consultant Published: July 2017 About SCMO 2 Founded in 2001, SCMO2 Specializes in High-End Supply Chain Consulting Work Focused

More information

Analytics for Banks. September 19, 2017

Analytics for Banks. September 19, 2017 Analytics for Banks September 19, 2017 Outline About AlgoAnalytics Problems we can solve for banks Our experience Technology Page 2 About AlgoAnalytics Analytics Consultancy Work at the intersection of

More information

Predictive compliance monitoring solution. Presented by The Red Flag Group

Predictive compliance monitoring solution. Presented by The Red Flag Group Predictive compliance monitoring solution Presented by The Red Flag Group We understand your compliance challenges Greater responsibility of compliance officers Regulative enforcements are increasingly

More information

No digitalization without risks

No digitalization without risks No digitalization without risks How to equip your organization against the new fraud threads? Frédéric Hennequin Senior Solution Specialist Fraud & Compliance Agenda SAS Introduction Online Fraud in Belgium

More information

Machine Learning 101

Machine Learning 101 Machine Learning 101 Mike Alperin September, 2016 Copyright 2000-2016 TIBCO Software Inc. Agenda What is Machine Learning? Decision Tree Models Customer Analytics Examples Manufacturing Examples Fraud

More information

Machine Learning Techniques For Particle Identification

Machine Learning Techniques For Particle Identification Machine Learning Techniques For Particle Identification 06.March.2018 I Waleed Esmail, Tobias Stockmanns, Michael Kunkel, James Ritman Institut für Kernphysik (IKP), Forschungszentrum Jülich Outlines:

More information

Enabling Foresight. Skills for Predictive Analytics

Enabling Foresight. Skills for Predictive Analytics Enabling Foresight Skills for Predictive Analytics Topic Outline INTRODUCTION Predictive Analytics Definition & Core Concepts Terms Used in Today s Analytics Environment Artificial Intelligence Deep Learning

More information

Rank hotels on Expedia.com to maximize purchases

Rank hotels on Expedia.com to maximize purchases Rank hotels on Expedia.com to maximize purchases Nishith Khantal, Valentina Kroshilina, Deepak Maini December 14, 2013 1 Introduction For an online travel agency (OTA), matching users to hotel inventory

More information

Traffic insights ATSPM Flashcards

Traffic insights ATSPM Flashcards Traffic insights ATSPM Flashcards Traffic Insights Optimize your traffic network. Measure the impact. Traffic Insights are a suite of signal performance analytics and tools that help improve traffic safety

More information

G U I D E B O O K P R O G R A M : D A T A A N D A N A L Y T I C S D O C U M E N T N U M B E R : S 180 N O V E M BER TENSORFLOW ON AWS

G U I D E B O O K P R O G R A M : D A T A A N D A N A L Y T I C S D O C U M E N T N U M B E R : S 180 N O V E M BER TENSORFLOW ON AWS 2 0 1 8 G U I D E B O O K P R O G R A M : D A T A A N D A N A L Y T I C S D O C U M E N T N U M B E R : S 180 N O V E M BER 2 0 1 8 TENSORFLOW ON AWS A N A L Y S T Rebecca Wettemann Nucleus Research, Inc.

More information

Using SAS Visual Investigator to Enforce Model Tuning Best Practices in a Regulatory Environment

Using SAS Visual Investigator to Enforce Model Tuning Best Practices in a Regulatory Environment Paper SAS1777-2018 Using SAS Visual Investigator to Enforce Model Tuning Best Practices in a Regulatory Environment Scott Wood, Josh Lincoln, and Edwin Rivera, SAS Institute Inc. ABSTRACT A common process

More information

Video Traffic Classification

Video Traffic Classification Video Traffic Classification A Machine Learning approach with Packet Based Features using Support Vector Machine Videotrafikklassificering En Maskininlärningslösning med Paketbasereade Features och Supportvektormaskin

More information

Large US Bank Boosts Insider Threat Detection by 5X with StreamAnalytix

Large US Bank Boosts Insider Threat Detection by 5X with StreamAnalytix Large US Bank Boosts Insider Threat Detection by 5X with StreamAnalytix About the customer About the Customer A large US-based financial services corporation known for its extensive credit card business

More information

TOKENIZATION: THE FUTURE OF ACCOUNT NUMBERS. Steve Ledford The Clearing House

TOKENIZATION: THE FUTURE OF ACCOUNT NUMBERS. Steve Ledford The Clearing House TOKENIZATION: THE FUTURE OF ACCOUNT NUMBERS Steve Ledford The Clearing House Problem Statement: The proliferation of live account credentials creates huge risks Bank issues physical card Plastic at point

More information

Credit Card Processing Guide

Credit Card Processing Guide Credit Card Processing Guide Detailed Overview of The Reports in The Envision Software I Credit Card Processing Within Envision Cloud Table of Contents Part I Introduction 1 Part II Processing On A Computer

More information

Paynalytix- As-a-Service. Accelerating Business Performance with Data-Driven Insights

Paynalytix- As-a-Service. Accelerating Business Performance with Data-Driven Insights Paynalytix- As-a-Service Accelerating Business Performance with Data-Driven Insights With the growth of digital payments, banks have a plethora of data available from an increasing number of digital payment

More information

A01 325: #1 VERSION 2 SOLUTIONS

A01 325: #1 VERSION 2 SOLUTIONS Economics 325: Public Economics Section A01 University of Victoria Midterm Examination #1 VERSION 2 SOLUTIONS Fall 2012 Instructor: Martin Farnham Midterm Exam #1 Section 1: Multiple Choice Each question

More information

Getting Started with Machine Learning

Getting Started with Machine Learning Getting Started with Machine Learning Machine learning is now easily accessible for companies of all sizes due to the recent availability of robust data platforms. And while these platforms make it faster

More information

Minimize Churn and Maximize Revenue. A Recurly Guide

Minimize Churn and Maximize Revenue. A Recurly Guide Minimize Churn and Maximize Revenue A Recurly Guide Introduction Churn is one of the key metrics in subscription commerce, tracked closely in niche start-ups and industry juggernauts alike. It is one of

More information

Kanban kick- start (v2)

Kanban kick- start (v2) Kanban kick- start (v2) By Tomas Björkholm at Crisp, October 2011 INTRODUCTION... 1 AN APPROACH TO GET STARTED WITH KANBAN... 2 STEP 1 GET TO KNOW YOUR SYSTEM... 2 STEP 2 IDENTIFY YOUR SOURCES AND PRIORITIZE...

More information

Introduction to Cognos Analytics and Report Navigation Training. IBM Cognos Analytics 11

Introduction to Cognos Analytics and Report Navigation Training. IBM Cognos Analytics 11 Introduction to Cognos Analytics and Report Navigation Training IBM Cognos Analytics 11 Applicable for former IBM Cognos 10 report users who access CBMS Cognos to run and view reports March 2018 This training

More information

Monetizing the Lake. Kirk Haslbeck, Hortonworks Dan Kernaghan, Pitney Bowes

Monetizing the Lake. Kirk Haslbeck, Hortonworks Dan Kernaghan, Pitney Bowes Monetizing the Lake Kirk Haslbeck, Hortonworks Dan Kernaghan, Pitney Bowes Hadoop is Lower Cost and more Scalable 14000 Cost Per Terabyte 12000 10000 8000 6000 4000 2000 0 HDP Oracle X Teradata Netezza

More information

My name is Sam Mulholland and I am the Managing Director of Standby Consulting.

My name is Sam Mulholland and I am the Managing Director of Standby Consulting. Cyber Security Forum Presentation C.B (Sam) Mulholland Good Afternoon My name is Sam Mulholland and I am the Managing Director of Standby Consulting. Just a little bit about myself. I have worked in IT

More information

Acknowledgments. Data Mining. Examples. Overview. Colleagues

Acknowledgments. Data Mining. Examples. Overview. Colleagues Data Mining A regression modeler s view on where it is and where it s likely to go. Bob Stine Department of Statistics The School of the Univ of Pennsylvania March 30, 2006 Acknowledgments Colleagues Dean

More information

A Bid for Every Auction. Implementing and Improving AdWords Automated Bidding

A Bid for Every Auction. Implementing and Improving AdWords Automated Bidding A Bid for Every Auction Implementing and Improving AdWords Automated Bidding Ten Automated Bidding Takeaways The Advantages of Auction-time Bidding 1 Bid to the user s context, as close to auction-time

More information

U.S. EMV Migration Update. A joint presentation from Citizens Commercial Banking and Worldpay

U.S. EMV Migration Update. A joint presentation from Citizens Commercial Banking and Worldpay U.S. EMV Migration Update A joint presentation from Citizens Commercial Banking and Worldpay Mike Phelps National Director Convenience & Retail Fuel Finance Holly Fengler Vice President National Customer

More information

PCI BLOG. P2PE, EMV, Tokenization, Oh My!

PCI BLOG. P2PE, EMV, Tokenization, Oh My! Page 1 of 8 PCI BLOG THE UNOFFICIAL PCI COMPLIANCE & IT SECURITY BLOG HOME PCI IN THE NEWS PCI TOOLS IT SEC. JOB BOARD DOCUMENTS CONTACT US FORUM P2PE, EMV, Tokenization, Oh My! June 14, 2016 PCI Blog

More information

LECTURE 17: MULTIVARIABLE REGRESSIONS I

LECTURE 17: MULTIVARIABLE REGRESSIONS I David Youngberg BSAD 210 Montgomery College LECTURE 17: MULTIVARIABLE REGRESSIONS I I. What Determines a House s Price? a. Open Data Set 6 to help us answer this question. You ll see pricing data for homes

More information

Using AI to Make Predictions on Stock Market

Using AI to Make Predictions on Stock Market Using AI to Make Predictions on Stock Market Alice Zheng Stanford University Stanford, CA 94305 alicezhy@stanford.edu Jack Jin Stanford University Stanford, CA 94305 jackjin@stanford.edu 1 Introduction

More information

Einführung in Amazon Machine Learning

Einführung in Amazon Machine Learning Einführung in Amazon Machine Learning Oliver Arafat Technical Evangelist Amazon Web Services arafato@amazon.de @OliverArafat Agenda Machine learning and the data ecosystem Smart applications by example

More information

What personal information do we collect, why, and what are your rights?

What personal information do we collect, why, and what are your rights? Privacy Policy This Privacy Policy lets you know what happens to any personal data that you provide to us, or any that may we collect about you. If you need any more, contact us at privacy@fanatics.co.uk.

More information

A Beginner s Guide to Machine Learning in Payment Fraud Detection P...

A Beginner s Guide to Machine Learning in Payment Fraud Detection P... 1 of 9 6/6/2018, 11:45 AM Nethone Sign up A Beginner's Guide to Machine Learning in Payment Fraud Detection & Prevention Aleksander Kijek, Chief Product Officer Aug 18 2017._ 1n While attending numerous

More information

Getting Started Quicken 2010 for Windows

Getting Started Quicken 2010 for Windows Getting Started Quicken 2010 for Windows Thank you for choosing Quicken 2010. This guide will get you up and running by stepping you through the key tasks you need to get started managing your finances

More information

Governance Grows Up as Outcomes Become Key

Governance Grows Up as Outcomes Become Key Governance Grows Up as Outcomes Become Key Outsourcing Center, Beth Ellyn Rosenthal, Senior Writer January 20, 2014 Ask 10 people what constitutes a successful long-term outsourcing relationship and you

More information

MANAGING FORWARD: Analytics for Today s Multi-Channel, Multi- Device Consumer. Larry Freed President &

MANAGING FORWARD: Analytics for Today s Multi-Channel, Multi- Device Consumer. Larry Freed President & MANAGING FORWARD: Analytics for Today s Multi-Channel, Multi- Device Consumer Larry Freed President & CEO @larryfreed larry.freed@foresee.com Managing Forward Using analytics to manage your business forward

More information

Is Your Company Ready for Digital Transformation? Introducing the Digital Readiness Index

Is Your Company Ready for Digital Transformation? Introducing the Digital Readiness Index Is Your Company Ready for Digital Transformation? Introducing the Digital Readiness Index Sight Machine regularly meets with large manufacturers who are embarking on journeys of digital transformation.

More information

Payment Strategies and Considerations for Transit. David L. dekozan Vice President, Strategic Initiatives Cubic Transportation Systems, Inc.

Payment Strategies and Considerations for Transit. David L. dekozan Vice President, Strategic Initiatives Cubic Transportation Systems, Inc. Payment Strategies and Considerations for Transit David L. dekozan Vice President, Strategic Initiatives Cubic Transportation Systems, Inc. Key Themes How did we get here? AFC 101 in 10 minutes or less

More information

8 Strategies FACEBOOK ADVERTISING. to Crush Your

8 Strategies FACEBOOK ADVERTISING. to Crush Your 8 Strategies to Crush Your FACEBOOK ADVERTISING PROCRACK TEAM 5000 + Collection Of AFFILIATE MARKETING Training Courses, WSO, Software's, Themeforest Themes, Codecanyan Scripts, Plugins & MUCH MORE...

More information

Credit and Debit Card Fraud

Credit and Debit Card Fraud Credit and Debit Card Fraud The Electronic Payment World, A Multi- Billion Dollar Market According to The Nilson Report in 2014 there were: US$89.93 Billion dollars in credit card transactions. US$105.63

More information

Drive Better Insights with Oracle Analytics Cloud

Drive Better Insights with Oracle Analytics Cloud Drive Better Insights with Oracle Analytics Cloud Thursday, April 5, 2018 Speakers: Jason Little, Sean Suskind Copyright 2018 Sierra-Cedar, Inc. All rights reserved Today s Presenters Jason Little VP of

More information

SAP Predictive Analytics Suite

SAP Predictive Analytics Suite SAP Predictive Analytics Suite Tania Pérez Asensio Where is the Evolution of Business Analytics Heading? Organizations Are Maturing Their Approaches to Solving Business Problems Reactive Wait until a problem

More information

Cognitive Data Governance

Cognitive Data Governance IBM Unified Governance & Integration White Paper Powered by Machine Learning to find and use governed data Jo Ramos Distinguished Engineer & Director IBM Analytics Rakesh Ranjan Program Director & Data

More information

q S pq S cq S. where q is the total amount produced/consumed. There are two representative firms that maximizes their profits, max pq 1 c 2 q2 1 q 1

q S pq S cq S. where q is the total amount produced/consumed. There are two representative firms that maximizes their profits, max pq 1 c 2 q2 1 q 1 Market design, problem set 2 (externalities, incomplete information, search) Do five of nine. 1. There is a representative consumer who picks a quantity q D to imize his utility, b log(q D ) pq D + w X

More information

Oracle Revenue Management and Billing Product Manager s Workbench. User Guide. Version Revision 1.1. E December, 2018

Oracle Revenue Management and Billing Product Manager s Workbench. User Guide. Version Revision 1.1. E December, 2018 Oracle Revenue Management and Billing Product Manager s Workbench Version 2.7.0.0.0 User Guide Revision 1.1 E88591-01 December, 2018 Copyright 2017, Oracle and/or its affiliates. All rights reserved. 1

More information

Online Algorithms and Competitive Analysis. Spring 2018

Online Algorithms and Competitive Analysis. Spring 2018 Online Algorithms and Competitive Analysis CS16: Introduction to Data Structures & Algorithms CS16: Introduction to Data Structures & Algorithms Spring 2018 Outline 1. Motivation 2. The Ski-Rental Problem

More information

The Mystery Behind Project Management Metrics. Reed Shell Blue Hippo Consulting

The Mystery Behind Project Management Metrics. Reed Shell Blue Hippo Consulting The Mystery Behind Project Management Metrics Reed Shell Blue Hippo Consulting Presentation Take-Aways Two Tools for gathering and producing metrics 10 Step Process Goal/Question/Metric Deliverable Exercises

More information

TT SCORE. Trade Surveillance with Machine Learning THE NEED FOR TRADE SURVEILLANCE TRADITIONAL PARAMETER-BASED SURVEILLANCE TOOLS

TT SCORE. Trade Surveillance with Machine Learning THE NEED FOR TRADE SURVEILLANCE TRADITIONAL PARAMETER-BASED SURVEILLANCE TOOLS TT SCORE Trade Surveillance with Machine Learning THE NEED FOR TRADE SURVEILLANCE The financial industry has experienced great change and extraordinary challenges in the wake of the 2008 global financial

More information

DATA SCIENCE OPS IN PRACTICE

DATA SCIENCE OPS IN PRACTICE DATA SCIENCE OPS IN PRACTICE Learn How Splunk Enables Fast Science for Cybersecurity Operations Innovation center, Washington, D.C. OLISA STEPHENSBAILEY DAVID BRENMAN SEPTEMBER 2017 DATA SCIENCE OPS IN

More information

TSYS Foresight ScoreSM with Featurespace : FAQs

TSYS Foresight ScoreSM with Featurespace : FAQs TSYS Foresight ScoreSM with Featurespace : FAQs What is Foresight Score is a fraud- and risk-management decisioning scoring tool that incorporates innovative machine learning capabilities to deliver significant

More information

FRAUD ANALYTICS AND DATA WAREHOUSING

FRAUD ANALYTICS AND DATA WAREHOUSING FRAUD ANALYTICS AND DATA WAREHOUSING Office of the Comptroller State of Maryland Andrew Schaufele; Director, Bureau of Revenue Estimates Agenda Our Team Background Case with TBD Success Successful Case

More information

It s Not Too Late for EMV What You Need To Do Now!

It s Not Too Late for EMV What You Need To Do Now! It s Not Too Late for EMV What You Need To Do Now! September, 2015 Lindsay Breathitt, Product Marketing Steve Cole, Product Management Ray Moorman, Product Management What is EMV? Fiction vs. Fact Why

More information

Automated Embedded AI Asset Intelligence. Jean-Michel Cambot Founder & Chief Evangelist

Automated Embedded AI Asset Intelligence. Jean-Michel Cambot Founder & Chief Evangelist Automated Embedded AI Asset Intelligence Jean-Michel Cambot Founder & Chief Evangelist Intelligent Machines must be able to explain every decision they make 2 Then comes the real Magic of Artificial Intelligence

More information

Local 2110 Instructions for Web Time Entry

Local 2110 Instructions for Web Time Entry Local 2110 Instructions for Web Time Entry Access MyTC Portal 1. Go to MyTC portal https://my.tc.columbia.edu/ 2. Enter UNI & Password Access Online Time Sheet 1. After accessing MyTC select the TC Services

More information