The Do s & Don ts of Building A Predictive Model in Insurance. University of Minnesota November 9 th, 2012 Nathan Hubbell, FCAS Katy Micek, Ph.D.

Save this PDF as:
 WORD  PNG  TXT  JPG

Size: px
Start display at page:

Download "The Do s & Don ts of Building A Predictive Model in Insurance. University of Minnesota November 9 th, 2012 Nathan Hubbell, FCAS Katy Micek, Ph.D."

Transcription

1 The Do s & Don ts of Building A Predictive Model in Insurance University of Minnesota November 9 th, 2012 Nathan Hubbell, FCAS Katy Micek, Ph.D.

2 Agenda Travelers Broad Overview Actuarial & Analytics Career Opportunities Actuarial & Analytics Leadership Development Program (AALDP) Do s and Don ts of Generalized Linear Models (GLMs) Insurance Background and Motivation Failings of Multiple Linear Regression Basics of GLMs Over-fitting Questions 2

3 About Travelers Offers property and casualty solutions to individuals and companies Second-largest commercial insurer in the U.S. Second-largest personal insurer through the independent agency channel Representatives in every U.S. state, Canada, Ireland and the United Kingdom A member of the Dow Jones Industrial Average the only insurance company on the list 3

4 Analytics at Travelers Who are we? Across Travelers, we form a large (300+) and diverse community of Ph.D., Masters and Bachelors holders in the following disciplines: mathematics statistics physics actuarial science computer science business and more! 4

5 Actuarial & Analytics Leadership Development Program (AALDP) 5-yr program for actuarial students and advanced analytics Actuarial track offers exam support Analytics learn insurance on the job through work projects and seminars Exams not required but support is provided for those interested Leadership development opportunities Career exploration opportunities through rotations Networking opportunities (mentor program, committees) 2012 Pilot Class for Advanced Analytics Offers a flexible career path 5

6 Generalized Linear Models 6

7 Basics of Insurance Insurance companies sell the product of insurance policies, which are the promise to pay in the event that a customer experiences a loss. The unique challenge in insurance is that we don t know what the cost of insuring a customer is when we sell the policy. Example: The cost to insure an auto customer It s impossible to predict if someone is going to Get into an accident The type of accident (hit a telephone pole, hit another vehicle, bodily injury) How bad (cost) the accident will be 7

8 Business Impact of Loss Experience To estimate the cost of insuring policyholders, we must predict losses Two fundamental questions we must answer are: 1. Ratemaking: looking to the future Setting rates for policies How much do we need to charge customers for a policy in order to reach our target profit? Basic idea: price = cost + profit 2. Reserving: looking at the impact of past experience Setting aside reserve money How much money do we need to set aside to pay for claims? Note: We cannot predict losses for each individual. However, if we group our customers together, we can build statistical models to predict average loss over a group. 8

9 First Model Attempt: Multiple Linear Regression / Ordinary Least Squares E[Y] = a 0 + a 1 X a n X n Goal: Fit a linear relationship between the predictors (X 1,, X n ) and the response variable Y. Y, the response Assumptions: 1. Y is normally distributed. 2. The variance of Y is constant. X, the predictor Approach: The parameters (a 0, a 1,, a n ) can be estimated by minimizing the sum of squared errors. 9

10 Oops - DON T assume Y is normally distributed In insurance, we study loss experience in terms of claims. Two aspects of claims must be considered. 1. Frequency: what is the rate that claims are being made? 2. Severity: what is the average size of claim? The underlying distribution in the model depends on what aspect of the loss experience we re investigating. 10

11 Double Oops - DON T assume the Variance of Y is constant Varies by expected loss. High frequency losses have less variance. High severity losses have more variance. 11

12 DO assume an exponential family distribution for Y Note: Non-normal distributions are more suited to highly skewed claim data Poisson - claim frequency Discrete distribution Time-invariant Variance equals mean (m = E[Y]) Gamma - severity Continuous distribution Variance equals mean squared (m 2 = E[Y] 2 ) Gamma Distribution Source: Gamma Distribution, Wikipedia 12

13 Suitable Model Framework: Generalized Linear Models (GLMs) E[Y] = g -1 (a 0 + a 1 X a n X n ) where g(x) is the link function. Goal: fit a non-linear relationship between the predictors (X 1,, X n ) and the response variable Y. Assumptions: 1. Y can be from any exponential family of distributions. 2. Variance depends on expected mean. Approach: The parameters (a 0, a 1,, a n ) can be estimated using maximum likelihood when underlying distribution is fixed.

14 Exponential Family ABCs Source: A Practitioner's Guide to Generalized Linear Models 14

15 Exponential Family Mean and Variance Tweedie V(m) = m p ; 1 < p < 2 Source: A Practitioner's Guide to Generalized Linear Models 15

16 Frequency & Severity All-in-One: Uncle Tweedie Source: A Practitioner's Guide to Generalized Linear Models 16

17 Over-fitting 17

18 DON T Underfit First let s start with under-fitting Expected Auto Claim Frequency: 5% Expected Auto Loss: $10,000 Expected Premium = 5% * $10,000 = $500 Competitors use a segmented rate plan What happens next? We will win all of the business where competitors charge more than $500 We lose all of the business where competitors charge less than $500 Now why would a competitor want to charge more than the average? Hmmm perhaps we need a better approach 18

19 Occam s Razor / Principle of Parsimony 19

20 However, DON T Overfit A noisy model = similar problems to under-fitting Some things you might use to look at fit: Summary Statistics (R 2, AIC, BIC) Deviance Tests / Type III P-value / Parameter Estimate Standard Errors ROC / Lift / Gain Charts However, let s keep our eyes on the prize: We want our model to perform well on future data! 20

21 DO Fit on Non-Training (Future!) Data Cross Validation Validation and Test / Hold out Bias / Variance Tradeoff Regularization 21

22 Cross Validation Training Validation Testing (Holdout) Model Fit Model Stucture Final Model Testing 22

23 Variance / Bias Tradeoff Source: Elements of Statistical Learning 23

24 Price Bias/variance: How would you fit this model? Size 24

25 Price Price Price Bias vs. Variance Size Size Size High bias (under-fit) Just right High variance (over-fit) Source: Coursera Machine Learning 25

26 Price Regularization Size High variance (overfit) Source: Coursera Machine Learning 26

27 References and Resources Actuarial Exams and Career Information Travelers Careers Actuarial and Analytics Research Internship and Full Time A Practitioner's Guide to Generalized Linear Models tion_3.pdf Statistical Modeling: The Two Cultures (Leo Breiman) Elements of Statistical Learning (Hastie, Tibshirani, Friedman) Coursera Machine Learning: 27

A Deeper Look Inside Generalized Linear Models

A Deeper Look Inside Generalized Linear Models A Deeper Look Inside Generalized Linear Models University of Minnesota February 3 rd, 2012 Nathan Hubbell, FCAS Agenda Property & Casualty (P&C Insurance) in one slide The Actuarial Profession Travelers

More information

Travelers Analytics: U of M Stats 8053 Insurance Modeling Problem

Travelers Analytics: U of M Stats 8053 Insurance Modeling Problem Travelers Analytics: U of M Stats 8053 Insurance Modeling Problem October 30 th, 2013 Nathan Hubbell, FCAS Shengde Liang, Ph.D. Agenda Travelers: Who Are We & How Do We Use Data? Insurance 101 Basic business

More information

Introduction to Predictive Modeling Using GLMs

Introduction to Predictive Modeling Using GLMs Introduction to Predictive Modeling Using GLMs Dan Tevet, FCAS, MAAA, Liberty Mutual Insurance Group Anand Khare, FCAS, MAAA, CPCU, Milliman 1 Antitrust Notice The Casualty Actuarial Society is committed

More information

GLM I An Introduction to Generalized Linear Models

GLM I An Introduction to Generalized Linear Models GLM I An Introduction to Generalized Linear Models CAS Ratemaking and Product Management Seminar March 2009 Presented by: Tanya D. Havlicek, Actuarial Assistant 0 ANTITRUST Notice The Casualty Actuarial

More information

BOOSTED REGRESSION TREES: A MODERN WAY TO ENHANCE ACTUARIAL MODELLING

BOOSTED REGRESSION TREES: A MODERN WAY TO ENHANCE ACTUARIAL MODELLING BOOSTED REGRESSION TREES: A MODERN WAY TO ENHANCE ACTUARIAL MODELLING Xavier Conort xavier.conort@gear-analytics.com Session Number: TBR14 Insurance has always been a data business The industry has successfully

More information

A new paradigm in P&C Industry Pricing

A new paradigm in P&C Industry Pricing The New Paradigm of Property & Casualty Insurance Pricing: Multivariate analysis and Predictive Modeling The ability to effectively price personal lines insurance policies to accurately match rate with

More information

We hope you will find this sample a useful supplement to your existing educational materials, and we look forward to receiving your comments.

We hope you will find this sample a useful supplement to your existing educational materials, and we look forward to receiving your comments. To Teachers of Mathematics: Thank you for visiting the BeAnActuary.org booth at the annual meeting of the National Council of Teachers of Mathematics. BeAnActuary.org is sponsored by the Joint Career Encouragement

More information

Mathematics of Risk. Introduction. Case Study #1 Personal Auto Insurance Pricing. Mathematical Concepts Illustrated. Background

Mathematics of Risk. Introduction. Case Study #1 Personal Auto Insurance Pricing. Mathematical Concepts Illustrated. Background Mathematics of Risk Introduction There are many mechanisms that individuals and organizations use to protect themselves against the risk of financial loss. Government organizations and public and private

More information

Combining Linear and Non-Linear Modeling Techniques: EMB America. Getting the Best of Two Worlds

Combining Linear and Non-Linear Modeling Techniques: EMB America. Getting the Best of Two Worlds Combining Linear and Non-Linear Modeling Techniques: Getting the Best of Two Worlds Outline Who is EMB? Insurance industry predictive modeling applications EMBLEM our GLM tool How we have used CART with

More information

Penalized regression: Introduction

Penalized regression: Introduction Penalized regression: Introduction Patrick Breheny August 30 Patrick Breheny BST 764: Applied Statistical Modeling 1/19 Maximum likelihood Much of 20th-century statistics dealt with maximum likelihood

More information

Offset Techniques for Predictive Modeling for Insurance

Offset Techniques for Predictive Modeling for Insurance Offset Techniques for Predictive Modeling for Insurance Matthew Flynn, Ph.D, ISO Innovative Analytics, W. Hartford CT Jun Yan, Ph.D, Deloitte & Touche LLP, Hartford CT ABSTRACT This paper presents the

More information

Model Validation Techniques

Model Validation Techniques Model Validation Techniques Kevin Mahoney, FCAS kmahoney@ travelers.com CAS RPM Seminar March 17, 2010 Uses of Statistical Models in P/C Insurance Examples of Applications Determine expected loss cost

More information

Risk pricing for Australian Motor Insurance

Risk pricing for Australian Motor Insurance Risk pricing for Australian Motor Insurance Dr Richard Brookes November 2012 Contents 1. Background Scope How many models? 2. Approach Data Variable filtering GLM Interactions Credibility overlay 3. Model

More information

Predictive Modeling Techniques in Insurance

Predictive Modeling Techniques in Insurance Predictive Modeling Techniques in Insurance Tuesday May 5, 2015 JF. Breton Application Engineer 2014 The MathWorks, Inc. 1 Opening Presenter: JF. Breton: 13 years of experience in predictive analytics

More information

Anti-Trust Notice. Agenda. Three-Level Pricing Architect. Personal Lines Pricing. Commercial Lines Pricing. Conclusions Q&A

Anti-Trust Notice. Agenda. Three-Level Pricing Architect. Personal Lines Pricing. Commercial Lines Pricing. Conclusions Q&A Achieving Optimal Insurance Pricing through Class Plan Rating and Underwriting Driven Pricing 2011 CAS Spring Annual Meeting Palm Beach, Florida by Beth Sweeney, FCAS, MAAA American Family Insurance Group

More information

GENERALIZED LINEAR MODELS IN VEHICLE INSURANCE

GENERALIZED LINEAR MODELS IN VEHICLE INSURANCE ACTA UNIVERSITATIS AGRICULTURAE ET SILVICULTURAE MENDELIANAE BRUNENSIS Volume 62 41 Number 2, 2014 http://dx.doi.org/10.11118/actaun201462020383 GENERALIZED LINEAR MODELS IN VEHICLE INSURANCE Silvie Kafková

More information

Data Mining Approaches to Modeling Insurance Risk. Dan Steinberg, Mikhail Golovnya, Scott Cardell. Salford Systems 2009

Data Mining Approaches to Modeling Insurance Risk. Dan Steinberg, Mikhail Golovnya, Scott Cardell. Salford Systems 2009 Data Mining Approaches to Modeling Insurance Risk Dan Steinberg, Mikhail Golovnya, Scott Cardell Salford Systems 2009 Overview of Topics Covered Examples in the Insurance Industry Predicting at the outset

More information

Logistic Regression (a type of Generalized Linear Model)

Logistic Regression (a type of Generalized Linear Model) Logistic Regression (a type of Generalized Linear Model) 1/36 Today Review of GLMs Logistic Regression 2/36 How do we find patterns in data? We begin with a model of how the world works We use our knowledge

More information

Predictive Modeling for Life Insurers

Predictive Modeling for Life Insurers Predictive Modeling for Life Insurers Application of Predictive Modeling Techniques in Measuring Policyholder Behavior in Variable Annuity Contracts April 30, 2010 Guillaume Briere-Giroux, FSA, MAAA, CFA

More information

business statistics using Excel OXFORD UNIVERSITY PRESS Glyn Davis & Branko Pecar

business statistics using Excel OXFORD UNIVERSITY PRESS Glyn Davis & Branko Pecar business statistics using Excel Glyn Davis & Branko Pecar OXFORD UNIVERSITY PRESS Detailed contents Introduction to Microsoft Excel 2003 Overview Learning Objectives 1.1 Introduction to Microsoft Excel

More information

Revising the ISO Commercial Auto Classification Plan. Anand Khare, FCAS, MAAA, CPCU

Revising the ISO Commercial Auto Classification Plan. Anand Khare, FCAS, MAAA, CPCU Revising the ISO Commercial Auto Classification Plan Anand Khare, FCAS, MAAA, CPCU 1 THE EXISTING PLAN 2 Context ISO Personal Lines Commercial Lines Property Homeowners, Dwelling Commercial Property, BOP,

More information

Measuring per-mile risk for pay-as-youdrive automobile insurance. Eric Minikel CAS Ratemaking & Product Management Seminar March 20, 2012

Measuring per-mile risk for pay-as-youdrive automobile insurance. Eric Minikel CAS Ratemaking & Product Management Seminar March 20, 2012 Measuring per-mile risk for pay-as-youdrive automobile insurance Eric Minikel CAS Ratemaking & Product Management Seminar March 20, 2012 Professor Joseph Ferreira, Jr. and Eric Minikel Measuring per-mile

More information

Addressing Analytics Challenges in the Insurance Industry. Noe Tuason California State Automobile Association

Addressing Analytics Challenges in the Insurance Industry. Noe Tuason California State Automobile Association Addressing Analytics Challenges in the Insurance Industry Noe Tuason California State Automobile Association Overview Two Challenges: 1. Identifying High/Medium Profit who are High/Low Risk of Flight Prospects

More information

Insurance Pricing Models Using Predictive Analytics. March 5 th, 2012

Insurance Pricing Models Using Predictive Analytics. March 5 th, 2012 Insurance Pricing Models Using Predictive Analytics March 5 th, 2012 2 Agenda Background/Information Why Predictive Analytics is Required Case Study-AMA An Effective Decision Tool for front-line Underwriting

More information

Staying Ahead of the Analytical Competitive Curve: Integrating the Broad Range Applications of Predictive Modeling in a Competitive Market Environment

Staying Ahead of the Analytical Competitive Curve: Integrating the Broad Range Applications of Predictive Modeling in a Competitive Market Environment Staying Ahead of the Analytical Competitive Curve: Integrating the Broad Range Applications of Predictive Modeling in a Competitive Market Environment Jun Yan, Ph.D., Mo Masud, and Cheng-sheng Peter Wu,

More information

Institute of Actuaries of India Subject CT3 Probability and Mathematical Statistics

Institute of Actuaries of India Subject CT3 Probability and Mathematical Statistics Institute of Actuaries of India Subject CT3 Probability and Mathematical Statistics For 2015 Examinations Aim The aim of the Probability and Mathematical Statistics subject is to provide a grounding in

More information

Invest in risk. A guide to a great career. The new General Insurance path to Fellowship with the Society of Actuaries.

Invest in risk. A guide to a great career. The new General Insurance path to Fellowship with the Society of Actuaries. A guide to a great career. The new General Insurance path to Fellowship with the Society of Actuaries. There will always be risk and uncertainty. Why not choose a profession dealing with it? The SOA is

More information

Actuarial. Modeling Seminar Part 2. Matthew Morton FSA, MAAA Ben Williams

Actuarial. Modeling Seminar Part 2. Matthew Morton FSA, MAAA Ben Williams Actuarial Data Analytics / Predictive Modeling Seminar Part 2 Matthew Morton FSA, MAAA Ben Williams Agenda Introduction Overview of Seminar Traditional Experience Study Traditional vs. Predictive Modeling

More information

Glossary of Insurance Terms

Glossary of Insurance Terms From the Insurance Bureau of Canada (IBC) Accident: An event that happens by chance and is not expected in the normal course of events, which results in harm to people, damage to property or equipment,

More information

Model selection in R featuring the lasso. Chris Franck LISA Short Course March 26, 2013

Model selection in R featuring the lasso. Chris Franck LISA Short Course March 26, 2013 Model selection in R featuring the lasso Chris Franck LISA Short Course March 26, 2013 Goals Overview of LISA Classic data example: prostate data (Stamey et. al) Brief review of regression and model selection.

More information

Predictive Modeling in Workers Compensation 2008 CAS Ratemaking Seminar

Predictive Modeling in Workers Compensation 2008 CAS Ratemaking Seminar Predictive Modeling in Workers Compensation 2008 CAS Ratemaking Seminar Prepared by Louise Francis, FCAS, MAAA Francis Analytics and Actuarial Data Mining, Inc. www.data-mines.com Louise.francis@data-mines.cm

More information

Predictive Modeling in Long-Term Care Insurance

Predictive Modeling in Long-Term Care Insurance Predictive Modeling in Long-Term Care Insurance Nathan R. Lally and Brian M. Hartman May 3, 2015 Abstract The accurate prediction of long-term care insurance (LTCI) mortality, lapse, and claim rates is

More information

Combining GLM and datamining techniques for modelling accident compensation data. Peter Mulquiney

Combining GLM and datamining techniques for modelling accident compensation data. Peter Mulquiney Combining GLM and datamining techniques for modelling accident compensation data Peter Mulquiney Introduction Accident compensation data exhibit features which complicate loss reserving and premium rate

More information

Predictive Analytics for Life Insurance: How Data and Advanced Analytics are Changing the Business of Life Insurance Seminar May 23, 2012

Predictive Analytics for Life Insurance: How Data and Advanced Analytics are Changing the Business of Life Insurance Seminar May 23, 2012 Predictive Analytics for Life Insurance: How Data and Advanced Analytics are Changing the Business of Life Insurance Seminar May 23, 2012 Session 2 How to Build a Risk Based Analytical Model for Life Insurance

More information

Multiple Linear Regression in Data Mining

Multiple Linear Regression in Data Mining Multiple Linear Regression in Data Mining Contents 2.1. A Review of Multiple Linear Regression 2.2. Illustration of the Regression Process 2.3. Subset Selection in Linear Regression 1 2 Chap. 2 Multiple

More information

Insurance Analytics - analýza dat a prediktivní modelování v pojišťovnictví. Pavel Kříž. Seminář z aktuárských věd MFF 4.

Insurance Analytics - analýza dat a prediktivní modelování v pojišťovnictví. Pavel Kříž. Seminář z aktuárských věd MFF 4. Insurance Analytics - analýza dat a prediktivní modelování v pojišťovnictví Pavel Kříž Seminář z aktuárských věd MFF 4. dubna 2014 Summary 1. Application areas of Insurance Analytics 2. Insurance Analytics

More information

Cross Validation techniques in R: A brief overview of some methods, packages, and functions for assessing prediction models.

Cross Validation techniques in R: A brief overview of some methods, packages, and functions for assessing prediction models. Cross Validation techniques in R: A brief overview of some methods, packages, and functions for assessing prediction models. Dr. Jon Starkweather, Research and Statistical Support consultant This month

More information

Cross Validation. Dr. Thomas Jensen Expedia.com

Cross Validation. Dr. Thomas Jensen Expedia.com Cross Validation Dr. Thomas Jensen Expedia.com About Me PhD from ETH Used to be a statistician at Link, now Senior Business Analyst at Expedia Manage a database with 720,000 Hotels that are not on contract

More information

Estimating Claim Settlement Values Using GLM. Roosevelt C. Mosley Jr., FCAS, MAAA

Estimating Claim Settlement Values Using GLM. Roosevelt C. Mosley Jr., FCAS, MAAA Estimating Claim Settlement Values Using GLM Roosevelt C. Mosley Jr., FCAS, MAAA 291 Estimating Claim Settlement Values Using GLM by Roosevelt C. Mosley, Jr., FCAS, MAAA Abstract: The goal of this paper

More information

Using Segmentation and Predictive Analytics to Combat Attrition

Using Segmentation and Predictive Analytics to Combat Attrition Using Segmentation and Predictive Analytics to Combat Attrition David Liebskind Retail Analytics Leader - GE Capital October 2011 1 Discussion Agenda 1 2 3 Understanding Attrition What is Attrition? How

More information

GLM III: Advanced Modeling Strategy 2005 CAS Seminar on Predictive Modeling Duncan Anderson MA FIA Watson Wyatt Worldwide

GLM III: Advanced Modeling Strategy 2005 CAS Seminar on Predictive Modeling Duncan Anderson MA FIA Watson Wyatt Worldwide GLM III: Advanced Modeling Strategy 25 CAS Seminar on Predictive Modeling Duncan Anderson MA FIA Watson Wyatt Worldwide W W W. W A T S O N W Y A T T. C O M Agenda Introduction Testing the link function

More information

Personal Auto Predictive Modeling Update: What s Next? Roosevelt Mosley, FCAS, MAAA CAS Predictive Modeling Seminar October 6 7, 2008 San Diego, CA

Personal Auto Predictive Modeling Update: What s Next? Roosevelt Mosley, FCAS, MAAA CAS Predictive Modeling Seminar October 6 7, 2008 San Diego, CA Personal Auto Predictive Modeling Update: What s Next? Roosevelt Mosley, FCAS, MAAA CAS Predictive Modeling Seminar October 6 7, 2008 San Diego, CA You ve Heard Where predictive modeling for auto has been

More information

Challenge. Solutions. Early results. Personal Lines Case Study Celina Insurance Reduces Expenses & Improves Processes Across the Business.

Challenge. Solutions. Early results. Personal Lines Case Study Celina Insurance Reduces Expenses & Improves Processes Across the Business. Celina Insurance Reduces Expenses & Improves Processes Across the Business About Celina Insurance Group Founded in 1914, Celina Insurance Group is composed of four mutual property and casualty insurance

More information

Better credit models benefit us all

Better credit models benefit us all Better credit models benefit us all Agenda Credit Scoring - Overview Random Forest - Overview Random Forest outperform logistic regression for credit scoring out of the box Interaction term hypothesis

More information

EDUCATION AND EXAMINATION COMMITTEE SOCIETY OF ACTUARIES RISK AND INSURANCE. Copyright 2005 by the Society of Actuaries

EDUCATION AND EXAMINATION COMMITTEE SOCIETY OF ACTUARIES RISK AND INSURANCE. Copyright 2005 by the Society of Actuaries EDUCATION AND EXAMINATION COMMITTEE OF THE SOCIET OF ACTUARIES RISK AND INSURANCE by Judy Feldman Anderson, FSA and Robert L. Brown, FSA Copyright 25 by the Society of Actuaries The Education and Examination

More information

Principle Component Analysis and Partial Least Squares: Two Dimension Reduction Techniques for Regression

Principle Component Analysis and Partial Least Squares: Two Dimension Reduction Techniques for Regression Principle Component Analysis and Partial Least Squares: Two Dimension Reduction Techniques for Regression Saikat Maitra and Jun Yan Abstract: Dimension reduction is one of the major tasks for multivariate

More information

AN INTRODUCTION TO PREMIUM TREND

AN INTRODUCTION TO PREMIUM TREND AN INTRODUCTION TO PREMIUM TREND Burt D. Jones * February, 2002 Acknowledgement I would like to acknowledge the valuable assistance of Catherine Taylor, who was instrumental in the development of this

More information

Predictive Modeling of Multi-Peril Homeowners. Insurance

Predictive Modeling of Multi-Peril Homeowners. Insurance Predictive Modeling of Multi-Peril Homeowners Insurance Edward W. (Jed) Frees Glenn Meyers A. David Cummings September 5, 2011 Abstract. Predictive models are used by insurers for underwriting and ratemaking

More information

Robert D. Helfand Carlton Fields Jorden Burt P.A. (860)

Robert D. Helfand Carlton Fields Jorden Burt P.A. (860) Robert D. Helfand Carlton Fields Jorden Burt P.A. rhelfand@carltonfields.com (860) 392-5044 P&C insurers expect big data use in many key business areas will more than double in the next two years. https://www.towerswatson.com/en-

More information

Predictive Modeling of Multi-Peril Homeowners. Insurance

Predictive Modeling of Multi-Peril Homeowners. Insurance Predictive Modeling of Multi-Peril Homeowners Insurance Edward W. (Jed) Frees Glenn Meyers A. David Cummings Abstract. Predictive models are used by insurers for underwriting and ratemaking in personal

More information

Example: Credit card default, we may be more interested in predicting the probabilty of a default than classifying individuals as default or not.

Example: Credit card default, we may be more interested in predicting the probabilty of a default than classifying individuals as default or not. Statistical Learning: Chapter 4 Classification 4.1 Introduction Supervised learning with a categorical (Qualitative) response Notation: - Feature vector X, - qualitative response Y, taking values in C

More information

Session 25 L, Introduction to General Insurance Ratemaking & Reserving: An Integrated Look. Moderator: W. Scott Lennox, FSA, FCAS, FCIA

Session 25 L, Introduction to General Insurance Ratemaking & Reserving: An Integrated Look. Moderator: W. Scott Lennox, FSA, FCAS, FCIA Session 25 L, Introduction to General Insurance Ratemaking & Reserving: An Integrated Look Moderator: W. Scott Lennox, FSA, FCAS, FCIA Presenter: Houston Cheng, FCAS, FCIA Society of Actuaries 2013 Annual

More information

PREDICTIVE LOSS RATIO

PREDICTIVE LOSS RATIO PREDICTIVE LOSS RATIO MODELING WITH CREDIT SCORES, FOR INSURANCE PURPOSES Major Qualifying Project submitted to the faculty of Worcester Polytechnic Institute in partial fulfillment of the requirements

More information

My name is Steven Lehmann. I am a Principal with Pinnacle Actuarial Resources, Inc., an actuarial consulting

My name is Steven Lehmann. I am a Principal with Pinnacle Actuarial Resources, Inc., an actuarial consulting Insurer Use of Education and Occupation Data National Conference of Insurance Legislators Special Property-Casualty Insurance Meeting February 28, 2009 My name is Steven Lehmann. I am a Principal with

More information

PREDICTIVE MODELLING FOR COMMERCIAL INSURANCE

PREDICTIVE MODELLING FOR COMMERCIAL INSURANCE PREDICTIVE MODELLING FOR COMMERCIAL INSURANCE General Insurance Pricing Seminar 13 June 2008 London James Guszcza, FCAS, MAAA jguszcza@deloitte.com General Themes Predictive modelling: 3 Levels of Discussion

More information

What can I do with a major in Business Administration?

What can I do with a major in Business Administration? Lewis-Clark State College offers a Bachelor of Arts or Science Degree in Business Administration through the Business Division. You can learn more about the Business Division and the Bachelor of Arts/Science

More information

Data Mining Techniques for Optimizing Québec s Automobile Risk-Sharing Pool

Data Mining Techniques for Optimizing Québec s Automobile Risk-Sharing Pool Data Mining Techniques for Optimizing Québec s Automobile Risk-Sharing Pool T E C H N O L O G I C A L W H I T E P A P E R Charles Dugas, Ph.D., A.S.A. Director, insurance solutions ApSTAT Technologies

More information

Final Report. Association of America Presentation

Final Report. Association of America Presentation Final Report 73 rd Conference of OK-AR Mathematical Association of America Presentation Principal Investigator: Marcel B. Finan Department of Mathematics Arkansas Tech University Russellville, AR 72801

More information

Actuarial Roundtable Presented by Milliman

Actuarial Roundtable Presented by Milliman Actuarial Roundtable Presented by Milliman Chicago RIMS Chapter October 21, 2014 Richard Frese, FCAS, MAAA Elizabeth Bart, FCAS, MAAA Mike Paczolt, FCAS, MAAA Tony Bloemer, FCAS, MAAA Doug Nishimura, ARM

More information

Generalised linear models for aggregate claims; to Tweedie or not?

Generalised linear models for aggregate claims; to Tweedie or not? Generalised linear models for aggregate claims; to Tweedie or not? Oscar Quijano and José Garrido Concordia University, Montreal, Canada (First version of May 20, 2013, this revision December 2, 2014)

More information

A Basic Guide to Modeling Techniques for All Direct Marketing Challenges

A Basic Guide to Modeling Techniques for All Direct Marketing Challenges A Basic Guide to Modeling Techniques for All Direct Marketing Challenges Allison Cornia Database Marketing Manager Microsoft Corporation C. Olivia Rud Executive Vice President Data Square, LLC Overview

More information

Statistical Learning Algorithms Applied to Automobile Insurance Ratemaking

Statistical Learning Algorithms Applied to Automobile Insurance Ratemaking Statistical Learning Algorithms Applied to Automobile Insurance Ratemaking C. Dugas, Y. Bengio, N. Chapados, P. Vincent, G. Denoncourt and C. Fournier The chapter will start from a description of the fundamentals

More information

Development Period 1 2 3 4 5 6 7 8 9 Observed Payments

Development Period 1 2 3 4 5 6 7 8 9 Observed Payments Pricing and reserving in the general insurance industry Solutions developed in The SAS System John Hansen & Christian Larsen, Larsen & Partners Ltd 1. Introduction The two business solutions presented

More information

Why do statisticians "hate" us?

Why do statisticians hate us? Why do statisticians "hate" us? David Hand, Heikki Mannila, Padhraic Smyth "Data mining is the analysis of (often large) observational data sets to find unsuspected relationships and to summarize the data

More information

Data Mining Practical Machine Learning Tools and Techniques

Data Mining Practical Machine Learning Tools and Techniques Ensemble learning Data Mining Practical Machine Learning Tools and Techniques Slides for Chapter 8 of Data Mining by I. H. Witten, E. Frank and M. A. Hall Combining multiple models Bagging The basic idea

More information

A Pricing Model for Underinsured Motorist Coverage

A Pricing Model for Underinsured Motorist Coverage A Pricing Model for Underinsured Motorist Coverage by Matthew Buchalter ABSTRACT Underinsured Motorist (UIM) coverage, also known as Family Protection coverage, is a component of most Canadian personal

More information

Some Essential Statistics The Lure of Statistics

Some Essential Statistics The Lure of Statistics Some Essential Statistics The Lure of Statistics Data Mining Techniques, by M.J.A. Berry and G.S Linoff, 2004 Statistics vs. Data Mining..lie, damn lie, and statistics mining data to support preconceived

More information

Introduction to Machine Learning. Speaker: Harry Chao Advisor: J.J. Ding Date: 1/27/2011

Introduction to Machine Learning. Speaker: Harry Chao Advisor: J.J. Ding Date: 1/27/2011 Introduction to Machine Learning Speaker: Harry Chao Advisor: J.J. Ding Date: 1/27/2011 1 Outline 1. What is machine learning? 2. The basic of machine learning 3. Principles and effects of machine learning

More information

More Flexible GLMs Zero-Inflated Models and Hybrid Models

More Flexible GLMs Zero-Inflated Models and Hybrid Models More Flexible GLMs Zero-Inflated Models and Hybrid Models Mathew Flynn, Ph.D. Louise A. Francis FCAS, MAAA Motivation: GLMs are widely used in insurance modeling applications. Claim or frequency models

More information

Umbrella Insurance: Sheltering You From a Sudden Storm UNCOMMON KNOWLEDGE. If you drive a car...you should own an umbrella policy.

Umbrella Insurance: Sheltering You From a Sudden Storm UNCOMMON KNOWLEDGE. If you drive a car...you should own an umbrella policy. UNCOMMON KNOWLEDGE Umbrella Insurance: Sheltering You From a Sudden Storm If you drive a car...you should own an umbrella policy. Consider this possibility... You are driving to work in your car, the same

More information

Mental Health Connection of Tarrant County

Mental Health Connection of Tarrant County Mental Health Connection of Tarrant County Health Insurance 101: Understanding the fundamentals of health insurance and the implications for health care reform Presented by: Tim Lee Fellow of Society of

More information

Math 370/408, Spring 2008 Prof. A.J. Hildebrand. Actuarial Exam Practice Problem Set 2 Solutions

Math 370/408, Spring 2008 Prof. A.J. Hildebrand. Actuarial Exam Practice Problem Set 2 Solutions Math 70/408, Spring 2008 Prof. A.J. Hildebrand Actuarial Exam Practice Problem Set 2 Solutions About this problem set: These are problems from Course /P actuarial exams that I have collected over the years,

More information

The Use of Education and Occupation as Underwriting Factors in Determining Policyholder Premiums for Private Passenger Auto Insurance

The Use of Education and Occupation as Underwriting Factors in Determining Policyholder Premiums for Private Passenger Auto Insurance The Use of Education and Occupation as Underwriting Factors in Determining Policyholder Premiums for Private Passenger Auto Insurance Written Testimony of Robert P. Hartwig, Ph.D., CPCU President & Chief

More information

33 DISCOUNTED PRESENT VALUE

33 DISCOUNTED PRESENT VALUE 33 DISCOUNTED PRESENT VALUE Purpose: To illustrate the idea of discounted present value with computations of the value of payments to be received in the future at different rates of interest. To use discounted

More information

Marketing Concept. The Marketing Concept

Marketing Concept. The Marketing Concept Marketing Concept Ted Mitchell is a philosophy of business competition. We need to know what it is, what it assumes, and what it implies. What is it? holds that the key to achieving organizational goals

More information

Emerging Leaders Program Business Operations Track

Emerging Leaders Program Business Operations Track Emerging Leaders Program Business Operations Track Locations: San Diego, Calif., Woodbury, NY, Buffalo, NY, Fredericksburg, Va., Virginia Beach, Va., Dallas, Texas, Lakeland, Fla., Macon, Ga., and Tucson,

More information

Insurance Fraud Detection: MARS versus Neural Networks?

Insurance Fraud Detection: MARS versus Neural Networks? Insurance Fraud Detection: MARS versus Neural Networks? Louise A Francis FCAS, MAAA Louise_francis@msn.com 1 Objectives Introduce a relatively new data mining method which can be used as an alternative

More information

Innovations and Value Creation in Predictive Modeling. David Cummings Vice President - Research

Innovations and Value Creation in Predictive Modeling. David Cummings Vice President - Research Innovations and Value Creation in Predictive Modeling David Cummings Vice President - Research ISO Innovative Analytics 1 Innovations and Value Creation in Predictive Modeling A look back at the past decade

More information

Classification and Regression Trees

Classification and Regression Trees Classification and Regression Trees Bob Stine Dept of Statistics, School University of Pennsylvania Trees Familiar metaphor Biology Decision tree Medical diagnosis Org chart Properties Recursive, partitioning

More information

Overall Level of Home Insurance Rates

Overall Level of Home Insurance Rates Overall Level of Home Insurance Rates University of Waterloo Actuarial Projects Competition Michael Vezina & Frédérick Guillot, Co-operators February 12 th, 2015 Agenda Who are we? Case Study Details?

More information

The zero-adjusted Inverse Gaussian distribution as a model for insurance claims

The zero-adjusted Inverse Gaussian distribution as a model for insurance claims The zero-adjusted Inverse Gaussian distribution as a model for insurance claims Gillian Heller 1, Mikis Stasinopoulos 2 and Bob Rigby 2 1 Dept of Statistics, Macquarie University, Sydney, Australia. email:

More information

Interaction Detection in GLM a Case Study

Interaction Detection in GLM a Case Study Interaction Detection in GLM a Case Study Chun Li, PhD ISO Innovative Analytics T H E S C I E N C E O F R I S K SM March 2012 1 Agenda Case study Approaches Proc Genmod, GAM in R, Proc Arbor Details Summary

More information

Data Mining Opportunities in Health Insurance

Data Mining Opportunities in Health Insurance Data Mining Opportunities in Health Insurance Methods Innovations and Case Studies Dan Steinberg, Ph.D. Copyright Salford Systems 2008 Analytical Challenges for Health Insurance Competitive pressures in

More information

Review Jeopardy. Blue vs. Orange. Review Jeopardy

Review Jeopardy. Blue vs. Orange. Review Jeopardy Review Jeopardy Blue vs. Orange Review Jeopardy Jeopardy Round Lectures 0-3 Jeopardy Round $200 How could I measure how far apart (i.e. how different) two observations, y 1 and y 2, are from each other?

More information

Find an expected value involving two events. Find an expected value involving multiple events. Use expected value to make investment decisions.

Find an expected value involving two events. Find an expected value involving multiple events. Use expected value to make investment decisions. 374 Chapter 8 The Mathematics of Likelihood 8.3 Expected Value Find an expected value involving two events. Find an expected value involving multiple events. Use expected value to make investment decisions.

More information

Decompose Error Rate into components, some of which can be measured on unlabeled data

Decompose Error Rate into components, some of which can be measured on unlabeled data Bias-Variance Theory Decompose Error Rate into components, some of which can be measured on unlabeled data Bias-Variance Decomposition for Regression Bias-Variance Decomposition for Classification Bias-Variance

More information

Session 62 TS, Predictive Modeling for Actuaries: Predictive Modeling Techniques in Insurance Moderator: Yonasan Schwartz, FSA, MAAA

Session 62 TS, Predictive Modeling for Actuaries: Predictive Modeling Techniques in Insurance Moderator: Yonasan Schwartz, FSA, MAAA Session 62 TS, Predictive Modeling for Actuaries: Predictive Modeling Techniques in Insurance Moderator: Yonasan Schwartz, FSA, MAAA Presenters: Jean-Frederic Breton David A. Moore, FSA, MAAA Session 62:

More information

Distinguishing the Forest from the TREES: A Comparison of Tree- Based Data Mining Methods

Distinguishing the Forest from the TREES: A Comparison of Tree- Based Data Mining Methods Distinguishing the Forest from the TREES: A Comparison of Tree- Based Data Mining Methods by Richard A. Derrig and Louise Francis ABSTRACT One of the most commonly used data mining techniques is decision

More information

Math 370, Actuarial Problemsolving Spring 2008 A.J. Hildebrand. Practice Test, 1/28/2008 (with solutions)

Math 370, Actuarial Problemsolving Spring 2008 A.J. Hildebrand. Practice Test, 1/28/2008 (with solutions) Math 370, Actuarial Problemsolving Spring 008 A.J. Hildebrand Practice Test, 1/8/008 (with solutions) About this test. This is a practice test made up of a random collection of 0 problems from past Course

More information

The Actuary and Street Pricing. General Insurance Actuaries. 03 November 2005

The Actuary and Street Pricing. General Insurance Actuaries. 03 November 2005 The Actuary and Street Pricing A dialogue to help you generate ideas to share views to list things that might affect our profession General Insurance Actuaries In the mid 70s some 30 actuaries, of which

More information

Service Management Capacity Planning and Queuing Models

Service Management Capacity Planning and Queuing Models Service Management Capacity Planning and Queuing Models Univ.-Prof. Dr.-Ing. Wolfgang Maass Chair in Economics Information and Service Systems (ISS) Saarland University, Saarbrücken, Germany WS 2011/2012

More information

RPM Workshop 3: Basic Ratemaking

RPM Workshop 3: Basic Ratemaking RPM Workshop 3: Basic Ratemaking Introduction to Ratemaking Relativities March 9, 2009 Mirage Hotel Las Vegas, NV Presented by: Chris Cooksey, FCAS, MAAA Nationwide Insurance Company Ain Milner, FCAS,

More information

Risk Analysis Overview

Risk Analysis Overview What Is Risk? Uncertainty about a situation can often indicate risk, which is the possibility of loss, damage, or any other undesirable event. Most people desire low risk, which would translate to a high

More information

Predictive Modeling of Multi-Peril Homeowners Insurance

Predictive Modeling of Multi-Peril Homeowners Insurance Predictive Modeling of Multi-Peril Homeowners Insurance by Edward W Frees, Glenn Meyers, and A David Cummings AbSTRACT Predictive models are used by insurers for underwriting and ratemaking in personal

More information

Predictive Analytics 101

Predictive Analytics 101 Predictive Analytics 101 Current Trends in Predictive Modeling and Analysis Frank A. Alerte, Esq.* *We would like to acknowledge the following individuals for their contribution and valuable input in preparing

More information

Predictive modelling around the world 28.11.13

Predictive modelling around the world 28.11.13 Predictive modelling around the world 28.11.13 Agenda Why this presentation is really interesting Introduction to predictive modelling Case studies Conclusions Why this presentation is really interesting

More information

Product Management Case Study. What makes Effective Product Management Process?

Product Management Case Study. What makes Effective Product Management Process? What makes Effective Product Management Process? The Product: The steps involved in developing a product to eventually manage What is involved in the development of a product? Identify a Niche Consumer

More information

Presentation to the LRC Automobile Insurance Modernization Committee

Presentation to the LRC Automobile Insurance Modernization Committee Presentation to the LRC Automobile Insurance Modernization Committee Rose Vaughn Williams December 6, 2011 In North Carolina, as in all states but a few, anyone who wants to drive on our roads is required

More information

STA 4273H: Statistical Machine Learning

STA 4273H: Statistical Machine Learning STA 4273H: Statistical Machine Learning Russ Salakhutdinov Department of Statistics! rsalakhu@utstat.toronto.edu! http://www.cs.toronto.edu/~rsalakhu/ Lecture 6 Three Approaches to Classification Construct

More information

Chapter 9 Experience rating

Chapter 9 Experience rating 0 INTRODUCTION 1 Chapter 9 Experience rating 0 Introduction The rating process is the process of deciding on an appropriate level of premium for a particular class of insurance business. The contents of

More information