Under the Basel II Accord, financial institutions are required for the first time to determine capital requirements for a new class of risk – operational risk. Large and internally active banks are required to estimate operational risk exposure using the Advanced Measurement Approach (AMA), which relies on advanced empirical models. As banks continue to develop and enhance their own AMA models for operational risk measurement, they are increasingly utilizing R to perform various modeling tasks.
In this presentation, Northern Trust will discuss the use of R in the loss distribution approach (LDA), the most widely used empirical approach for the measurement of operational risk. In Northern Trust’s experience, R offers unparalleled access to various distributions that are most relevant for modeling the frequency and severity of operational loss events. Additionally, Northern Trust utilizes R to perform large scale Monte Carlo simulations within the context of the LDA. These simulations are computationally intense from a processing perspective, taking many hours and sometimes days to complete with the open source distribution but much less with Revolution R Enterprise.
How to Drive Value from Operational Risk Data - Part 2Perficient, Inc.
As complexities in the financial markets continue to increase, so too does the challenge of understanding and mitigating operational risks that can negatively affect the business. Many firms still struggle with risk identification and how data can be leveraged across the enterprise to prevent operational risk losses and gain operational efficiencies.
During this webinar, Perficient’s industry experts discussed the evolving role and challenges of operational risk management (ORM) in financial services, tips for a comprehensive approach to identify, assess and mitigate risks, and strategies to gain value from operational risk data to support the business.
Operational risk management and measurementRahmat Mulyana
a short description in mixed English and Bahasa Indonesia on Operational Risk Management and Measurement, in particular value at risk calculation using Monte carlo Simulation. Another method using EVT (Extree Value Theory) will be delivered shortly. regards
The document discusses supply chain risk management and minimizing risk exposure. It outlines various risks in the supply chain from external factors like the environment and demand as well as internal factors like processes and governance. It emphasizes the need for a risk framework that includes strategy, execution, and continuous improvement. Key aspects of risk management include risk planning, managing suppliers and inventory, and having the right competencies and performance metrics.
Presented to eRum (Budapest), May 2018
There are many common workloads in R that are "embarrassingly parallel": group-by analyses, simulations, and cross-validation of models are just a few examples. In this talk I'll describe the doAzureParallel package, a backend to the "foreach" package that automates the process of spawning a cluster of virtual machines in the Azure cloud to process iterations in parallel. This will include an example of optimizing hyperparameters for a predictive model using the "caret" package.
By David Smith. Presented at Microsoft Build (Seattle), May 7 2018.
Your data scientists have created predictive models using open-source tools, proprietary software, or some combination of both, and now you are interested in lifting and shifting those models to the cloud. In this talk, I'll describe how data scientists can transition their existing workflows — while using mostly the same tools and processes — to train and deploy machine learning models based on open source frameworks to Azure. I'll provide guidance on keeping connections to data sources up-to-date, evaluating and monitoring models, and deploying applications that make use of those models.
Presentation delivered by David Smith to NY R Conference https://www.rstats.nyc/, April 2018:
Minecraft is an open-world creativity game, and a hit with kids. To get kids interested in learning to program with R, we created the "miner" package. This package is a collection of simple functions that allow you to connect with a Minecraft instance, manipulate the world within by creating blocks and controlling the player, and to detect events within the world and react accordingly.
The miner package is intended mainly for kids, to inspire them to learn R while playing Minecraft. But the development of the package also provides some useful insights into how to build an R package to interface with a persistent API, and how to instruct others on its use. In this talk I'll describe how to set up your own Minecraft server, and how to use and extend the package. I'll also provide a few examples of the package in action in a live Minecraft session.
This short document appears to be discussing a vision API and mentions @revodavid but does not provide enough contextual information to generate a meaningful 3 sentence summary. The document is only two tweets with limited semantic content.
There are many common workloads in R that are "embarrassingly parallel": group-by analyses, simulations, and cross-validation of models are just a few examples. In this talk I'll describe several techniques available in R to speed up workloads like these, by running multiple iterations simultaneously, in parallel.
Many of these techniques require the use of a cluster of machines running R, and I'll provide examples of using cloud-based services to provision clusters for parallel computations. In particular, I will describe how you can use the SparklyR package to distribute data manipulations using the dplyr syntax, on a cluster of servers provisioned in the Azure cloud.
Presented by David Smith at Data Day Texas in Austin, January 27 2018.
How to Drive Value from Operational Risk Data - Part 2Perficient, Inc.
As complexities in the financial markets continue to increase, so too does the challenge of understanding and mitigating operational risks that can negatively affect the business. Many firms still struggle with risk identification and how data can be leveraged across the enterprise to prevent operational risk losses and gain operational efficiencies.
During this webinar, Perficient’s industry experts discussed the evolving role and challenges of operational risk management (ORM) in financial services, tips for a comprehensive approach to identify, assess and mitigate risks, and strategies to gain value from operational risk data to support the business.
Operational risk management and measurementRahmat Mulyana
a short description in mixed English and Bahasa Indonesia on Operational Risk Management and Measurement, in particular value at risk calculation using Monte carlo Simulation. Another method using EVT (Extree Value Theory) will be delivered shortly. regards
The document discusses supply chain risk management and minimizing risk exposure. It outlines various risks in the supply chain from external factors like the environment and demand as well as internal factors like processes and governance. It emphasizes the need for a risk framework that includes strategy, execution, and continuous improvement. Key aspects of risk management include risk planning, managing suppliers and inventory, and having the right competencies and performance metrics.
Presented to eRum (Budapest), May 2018
There are many common workloads in R that are "embarrassingly parallel": group-by analyses, simulations, and cross-validation of models are just a few examples. In this talk I'll describe the doAzureParallel package, a backend to the "foreach" package that automates the process of spawning a cluster of virtual machines in the Azure cloud to process iterations in parallel. This will include an example of optimizing hyperparameters for a predictive model using the "caret" package.
By David Smith. Presented at Microsoft Build (Seattle), May 7 2018.
Your data scientists have created predictive models using open-source tools, proprietary software, or some combination of both, and now you are interested in lifting and shifting those models to the cloud. In this talk, I'll describe how data scientists can transition their existing workflows — while using mostly the same tools and processes — to train and deploy machine learning models based on open source frameworks to Azure. I'll provide guidance on keeping connections to data sources up-to-date, evaluating and monitoring models, and deploying applications that make use of those models.
Presentation delivered by David Smith to NY R Conference https://www.rstats.nyc/, April 2018:
Minecraft is an open-world creativity game, and a hit with kids. To get kids interested in learning to program with R, we created the "miner" package. This package is a collection of simple functions that allow you to connect with a Minecraft instance, manipulate the world within by creating blocks and controlling the player, and to detect events within the world and react accordingly.
The miner package is intended mainly for kids, to inspire them to learn R while playing Minecraft. But the development of the package also provides some useful insights into how to build an R package to interface with a persistent API, and how to instruct others on its use. In this talk I'll describe how to set up your own Minecraft server, and how to use and extend the package. I'll also provide a few examples of the package in action in a live Minecraft session.
This short document appears to be discussing a vision API and mentions @revodavid but does not provide enough contextual information to generate a meaningful 3 sentence summary. The document is only two tweets with limited semantic content.
There are many common workloads in R that are "embarrassingly parallel": group-by analyses, simulations, and cross-validation of models are just a few examples. In this talk I'll describe several techniques available in R to speed up workloads like these, by running multiple iterations simultaneously, in parallel.
Many of these techniques require the use of a cluster of machines running R, and I'll provide examples of using cloud-based services to provision clusters for parallel computations. In particular, I will describe how you can use the SparklyR package to distribute data manipulations using the dplyr syntax, on a cluster of servers provisioned in the Azure cloud.
Presented by David Smith at Data Day Texas in Austin, January 27 2018.
The R Ecosystem consists of the R Foundation which oversees the R programming language and its core development. The R Core Group maintains the R software and CRAN which distributes R packages. A large contributor and user community provides documentation, blogs, user groups and additional software/services to support the widespread use of R.
A look at the changing perceptions of R, from the early days of the R project to today. Microsoft sponsor talk, presented by David Smith to the useR!2017 conference in Brussels, July 5 2017.
Predicting Loan Delinquency at One Million Transactions per SecondRevolution Analytics
Real-time applications of predictive models must be able to generate predictions at the rate that transactions are generated. Previously, such applications of models trained using R needed to be converted to other languages like C++ or Java to achieve the required throughput. In this talk, I’ll describe how to use the in-database R processing capabilities of Microsoft R Server to detect fraud in a SQL Server database of loan records at a rate exceeding one million transactions per second. I will also show the process of training the underlying gradient-boosted tree model on a large training set using the out-of-memory algorithms of Microsoft R.
Presented by David Smith at The Data Science Summit, Chicago, April 20 2017.
The ability to independently reproduce results is a critical issue within the scientific community today, and is equally important for collaboration and compliance in business. In this talk, I'll introduce several features available in R that help you make reproducibility a standard part of your data science workflow. The talk will include tips on working with data and files, combining code and output, and managing R's changing package ecosystem.
Presented by David Smith, R Community Lead (Microsoft), at Monktoberfest October 2016.
The value of open source isn’t just in the software itself. The communities that form around open source software provide just as much value and sometimes even more: in ongoing development, in documentation, in support, in marketing, and as a supply of ready-trained employees. Companies who build on open source tend to focus on the software, but neglect communities at their peril.
In this talk, I share some of my experiences in building community for an open-source software company, Revolution Analytics, and perspectives since the acquisition by Microsoft in 2015.
R is more than just a language. Many of the reasons why R has become such a popular tool for data science come from the ecosystem surrounding the R project. R users benefit from the many resources and packages created by the community, while commercial companies (including Microsoft) provide tools to extend and support R, and services to help people use R.
In this talk, I will give an overview of the R Ecosystem and describe how it has been a critical component of R’s success, and include several examples of Microsoft’s contributions to the ecosystem.
(Presented to EARL London, September 2016)
(Presented by David Smith at useR!2016, June 2016. Recording: https://channel9.msdn.com/Events/useR-international-R-User-conference/useR2016/R-at-Microsoft )
Since the acquisition of Revolution Analytics in April 2015, Microsoft has embarked upon a project to build R technology into many Microsoft products, so that developers and data scientists can use the R language and R packages to analyze data in their data centers and in cloud environments.
In this talk I will give an overview (and a demo or two) of how R has been integrated into various Microsoft products. Microsoft data scientists are also big users of R, and I'll describe a couple of examples of R being used to analyze operational data at Microsoft. I'll also share some of my experiences in working with open source projects at Microsoft, and my thoughts on how Microsoft works with open source communities including the R Project.
Hadoop is famously scalable. Cloud Computing is famously scalable. R – the thriving and extensible open source Data Science software – not so much. But what if we seamlessly combined Hadoop, Cloud Computing, and R to create a scalable Data Science platform? Imagine exploring, transforming, modeling, and scoring data at any scale from the comfort of your favorite R environment. Now, imagine calling a simple R function to operationalize your predictive model as a scalable, cloud-based Web Service. Learn how to leverage the magic of Hadoop on-premises or in the cloud to run your R code, thousands of open source R extension packages, and distributed implementations of the most popular machine learning algorithms at scale.
The document discusses Revolution Analytics, a company that provides analytics software and services based on the open source R language. It was acquired by Microsoft to help customers use advanced analytics within Microsoft data platforms. The document provides overviews of R, data science in the cloud using Azure, connecting R to SQL, solving scalability issues with Revolution R Enterprise (RRE), using R in SQL Server, and moving analytics workflows to the cloud.
The document compares two biological networks, CRAN and BioConductor, in terms of their topological properties. It finds that CRAN has more nodes and edges than BioConductor, but BioConductor is more densely connected. A statistical test shows the networks have significantly different degree distributions, though both approximately follow power laws.
This document discusses using graph analysis and PageRank algorithms to analyze the network structure and popularity of R packages over time. It contains the steps to create a dependency graph from package metadata, plot and export the graph, and shows the top 10 most popular packages according to PageRank in 2012 and 2015. Further analysis of the network structure is suggested.
Checkpoint provides a simple way to ensure reproducible results in R. It works by adding two lines to an R script that install the checkpoint package and specify a date. Checkpoint will then install all packages used in the script to the versions that were available on that date. This allows sharing of R code and ensuring others can reproduce results even if package versions change later. Checkpoint manages dependencies and installation of correct package versions for reproducibility across systems and time.
This document summarizes R at Microsoft, including its acquisition of Revolution Analytics. Key points include:
- R is a widely used open-source statistical programming language. Microsoft acquired Revolution Analytics to help customers use advanced analytics within its data platforms.
- Microsoft products like SQL Server will integrate Revolution R Open, an enhanced open-source R distribution, to allow running R scripts directly from SQL queries.
- Microsoft aims to make R and advanced analytics more accessible and scalable through products like the Machine Learning marketplace and by running R on servers to handle large datasets within SQL Server and Azure.
Revolution R Enterprise 7.4 - Presentation by Bill Jacobs 11Jun15Revolution Analytics
This document outlines several improvements and updates to ScaleR including new capabilities for DeployR, an upgraded R Engine, improved performance for various models, added support for HDFS caching and updated security features. It also notes changes to packages, platforms, and the separate installation of the R Engine.
With rising business challenges in the aftermarket service areas, it becomes imperative for manufacturers to gain actionable intelligence across the warranty management life cycle.
Join Revolution Analytics and Tech Mahindra to hear how to reduce the information visibility gap:
• Identify statistically significant business drivers
• Forecast warranty costs and claims
• Improve Customer Satisfaction
Presented by Joseph Rickert at the NYC R Conference, April 25 2015.
Good data analysis is reproducible. If someone else can’t independently replicate your results from your data, the consequences can be severe. With R, a major challenge for reproducibility is the ever-changing package ecosystem: it's all too easy to develop an R script using packages, only to find collaborators will download later versions of those packages when they attempt to reproduce your results, and outcome can be unpredictable!
In this talk I'll introduce the Reproducible R Toolkit, and the "checkpoint" package, included with Revolution R Open, and describe some best practices for writing reliable, reproducible R code with packages.
Reproducibility with Revolution R Open and the Checkpoint PackageRevolution Analytics
This document summarizes a webinar about reproducibility in R using the checkpoint package. It discusses how checkpoint allows users to specify a date that locks down package versions, ensuring results can be reproduced even if packages are later updated. The webinar demonstrates checkpoint's use via an example analyzing weather data. It promotes Revolution R Open, which includes checkpoint, for reproducible and shareable R code.
Performance and Scale Options for R with Hadoop: A comparison of potential ar...Revolution Analytics
R and Hadoop go together. In fact, they go together so well, that the number of options available can be confusing to IT and data science teams seeking solutions under varying performance and operational requirements.
Which configuration is faster for big files? Which is faster for sharing data and servers among groups? Which eliminates data movement? Which is easiest to manage? Which works best with iterative and multistep algorithms? What are the hardware requirements of each alternative?
This webinar is intended to help new users of R with Hadoop select their best architecture for integrating Hadoop and R, by explaining the benefits of several popular configurations, their performance potential, workload handling and programming model and administrative characteristics.
Presenters from Revolution Analytics will describe the options for using Revolution R Open and Revolution R Enterprise with Hadoop including servers, edge nodes, rHadoop and ScaleR. We’ll then compare the characteristics of each configuration as regards performance but also programming model, administration, data movement, ease of scaling, mixed workload handling, and performance for large individual analyses vs. mixed workloads.
Discover the Future of Dogecoin with Our Comprehensive Guidance36 Crypto
Learn in-depth about Dogecoin's trajectory and stay informed with 36crypto's essential and up-to-date information about the crypto space.
Our presentation delves into Dogecoin's potential future, exploring whether it's destined to skyrocket to the moon or face a downward spiral. In addition, it highlights invaluable insights. Don't miss out on this opportunity to enhance your crypto understanding!
https://36crypto.com/the-future-of-dogecoin-how-high-can-this-cryptocurrency-reach/
The R Ecosystem consists of the R Foundation which oversees the R programming language and its core development. The R Core Group maintains the R software and CRAN which distributes R packages. A large contributor and user community provides documentation, blogs, user groups and additional software/services to support the widespread use of R.
A look at the changing perceptions of R, from the early days of the R project to today. Microsoft sponsor talk, presented by David Smith to the useR!2017 conference in Brussels, July 5 2017.
Predicting Loan Delinquency at One Million Transactions per SecondRevolution Analytics
Real-time applications of predictive models must be able to generate predictions at the rate that transactions are generated. Previously, such applications of models trained using R needed to be converted to other languages like C++ or Java to achieve the required throughput. In this talk, I’ll describe how to use the in-database R processing capabilities of Microsoft R Server to detect fraud in a SQL Server database of loan records at a rate exceeding one million transactions per second. I will also show the process of training the underlying gradient-boosted tree model on a large training set using the out-of-memory algorithms of Microsoft R.
Presented by David Smith at The Data Science Summit, Chicago, April 20 2017.
The ability to independently reproduce results is a critical issue within the scientific community today, and is equally important for collaboration and compliance in business. In this talk, I'll introduce several features available in R that help you make reproducibility a standard part of your data science workflow. The talk will include tips on working with data and files, combining code and output, and managing R's changing package ecosystem.
Presented by David Smith, R Community Lead (Microsoft), at Monktoberfest October 2016.
The value of open source isn’t just in the software itself. The communities that form around open source software provide just as much value and sometimes even more: in ongoing development, in documentation, in support, in marketing, and as a supply of ready-trained employees. Companies who build on open source tend to focus on the software, but neglect communities at their peril.
In this talk, I share some of my experiences in building community for an open-source software company, Revolution Analytics, and perspectives since the acquisition by Microsoft in 2015.
R is more than just a language. Many of the reasons why R has become such a popular tool for data science come from the ecosystem surrounding the R project. R users benefit from the many resources and packages created by the community, while commercial companies (including Microsoft) provide tools to extend and support R, and services to help people use R.
In this talk, I will give an overview of the R Ecosystem and describe how it has been a critical component of R’s success, and include several examples of Microsoft’s contributions to the ecosystem.
(Presented to EARL London, September 2016)
(Presented by David Smith at useR!2016, June 2016. Recording: https://channel9.msdn.com/Events/useR-international-R-User-conference/useR2016/R-at-Microsoft )
Since the acquisition of Revolution Analytics in April 2015, Microsoft has embarked upon a project to build R technology into many Microsoft products, so that developers and data scientists can use the R language and R packages to analyze data in their data centers and in cloud environments.
In this talk I will give an overview (and a demo or two) of how R has been integrated into various Microsoft products. Microsoft data scientists are also big users of R, and I'll describe a couple of examples of R being used to analyze operational data at Microsoft. I'll also share some of my experiences in working with open source projects at Microsoft, and my thoughts on how Microsoft works with open source communities including the R Project.
Hadoop is famously scalable. Cloud Computing is famously scalable. R – the thriving and extensible open source Data Science software – not so much. But what if we seamlessly combined Hadoop, Cloud Computing, and R to create a scalable Data Science platform? Imagine exploring, transforming, modeling, and scoring data at any scale from the comfort of your favorite R environment. Now, imagine calling a simple R function to operationalize your predictive model as a scalable, cloud-based Web Service. Learn how to leverage the magic of Hadoop on-premises or in the cloud to run your R code, thousands of open source R extension packages, and distributed implementations of the most popular machine learning algorithms at scale.
The document discusses Revolution Analytics, a company that provides analytics software and services based on the open source R language. It was acquired by Microsoft to help customers use advanced analytics within Microsoft data platforms. The document provides overviews of R, data science in the cloud using Azure, connecting R to SQL, solving scalability issues with Revolution R Enterprise (RRE), using R in SQL Server, and moving analytics workflows to the cloud.
The document compares two biological networks, CRAN and BioConductor, in terms of their topological properties. It finds that CRAN has more nodes and edges than BioConductor, but BioConductor is more densely connected. A statistical test shows the networks have significantly different degree distributions, though both approximately follow power laws.
This document discusses using graph analysis and PageRank algorithms to analyze the network structure and popularity of R packages over time. It contains the steps to create a dependency graph from package metadata, plot and export the graph, and shows the top 10 most popular packages according to PageRank in 2012 and 2015. Further analysis of the network structure is suggested.
Checkpoint provides a simple way to ensure reproducible results in R. It works by adding two lines to an R script that install the checkpoint package and specify a date. Checkpoint will then install all packages used in the script to the versions that were available on that date. This allows sharing of R code and ensuring others can reproduce results even if package versions change later. Checkpoint manages dependencies and installation of correct package versions for reproducibility across systems and time.
This document summarizes R at Microsoft, including its acquisition of Revolution Analytics. Key points include:
- R is a widely used open-source statistical programming language. Microsoft acquired Revolution Analytics to help customers use advanced analytics within its data platforms.
- Microsoft products like SQL Server will integrate Revolution R Open, an enhanced open-source R distribution, to allow running R scripts directly from SQL queries.
- Microsoft aims to make R and advanced analytics more accessible and scalable through products like the Machine Learning marketplace and by running R on servers to handle large datasets within SQL Server and Azure.
Revolution R Enterprise 7.4 - Presentation by Bill Jacobs 11Jun15Revolution Analytics
This document outlines several improvements and updates to ScaleR including new capabilities for DeployR, an upgraded R Engine, improved performance for various models, added support for HDFS caching and updated security features. It also notes changes to packages, platforms, and the separate installation of the R Engine.
With rising business challenges in the aftermarket service areas, it becomes imperative for manufacturers to gain actionable intelligence across the warranty management life cycle.
Join Revolution Analytics and Tech Mahindra to hear how to reduce the information visibility gap:
• Identify statistically significant business drivers
• Forecast warranty costs and claims
• Improve Customer Satisfaction
Presented by Joseph Rickert at the NYC R Conference, April 25 2015.
Good data analysis is reproducible. If someone else can’t independently replicate your results from your data, the consequences can be severe. With R, a major challenge for reproducibility is the ever-changing package ecosystem: it's all too easy to develop an R script using packages, only to find collaborators will download later versions of those packages when they attempt to reproduce your results, and outcome can be unpredictable!
In this talk I'll introduce the Reproducible R Toolkit, and the "checkpoint" package, included with Revolution R Open, and describe some best practices for writing reliable, reproducible R code with packages.
Reproducibility with Revolution R Open and the Checkpoint PackageRevolution Analytics
This document summarizes a webinar about reproducibility in R using the checkpoint package. It discusses how checkpoint allows users to specify a date that locks down package versions, ensuring results can be reproduced even if packages are later updated. The webinar demonstrates checkpoint's use via an example analyzing weather data. It promotes Revolution R Open, which includes checkpoint, for reproducible and shareable R code.
Performance and Scale Options for R with Hadoop: A comparison of potential ar...Revolution Analytics
R and Hadoop go together. In fact, they go together so well, that the number of options available can be confusing to IT and data science teams seeking solutions under varying performance and operational requirements.
Which configuration is faster for big files? Which is faster for sharing data and servers among groups? Which eliminates data movement? Which is easiest to manage? Which works best with iterative and multistep algorithms? What are the hardware requirements of each alternative?
This webinar is intended to help new users of R with Hadoop select their best architecture for integrating Hadoop and R, by explaining the benefits of several popular configurations, their performance potential, workload handling and programming model and administrative characteristics.
Presenters from Revolution Analytics will describe the options for using Revolution R Open and Revolution R Enterprise with Hadoop including servers, edge nodes, rHadoop and ScaleR. We’ll then compare the characteristics of each configuration as regards performance but also programming model, administration, data movement, ease of scaling, mixed workload handling, and performance for large individual analyses vs. mixed workloads.
Discover the Future of Dogecoin with Our Comprehensive Guidance36 Crypto
Learn in-depth about Dogecoin's trajectory and stay informed with 36crypto's essential and up-to-date information about the crypto space.
Our presentation delves into Dogecoin's potential future, exploring whether it's destined to skyrocket to the moon or face a downward spiral. In addition, it highlights invaluable insights. Don't miss out on this opportunity to enhance your crypto understanding!
https://36crypto.com/the-future-of-dogecoin-how-high-can-this-cryptocurrency-reach/
Abhay Bhutada, the Managing Director of Poonawalla Fincorp Limited, is an accomplished leader with over 15 years of experience in commercial and retail lending. A Qualified Chartered Accountant, he has been pivotal in leveraging technology to enhance financial services. Starting his career at Bank of India, he later founded TAB Capital Limited and co-founded Poonawalla Finance Private Limited, emphasizing digital lending. Under his leadership, Poonawalla Fincorp achieved a 'AAA' credit rating, integrating acquisitions and emphasizing corporate governance. Actively involved in industry forums and CSR initiatives, Abhay has been recognized with awards like "Young Entrepreneur of India 2017" and "40 under 40 Most Influential Leader for 2020-21." Personally, he values mindfulness, enjoys gardening, yoga, and sees every day as an opportunity for growth and improvement.
OJP data from firms like Vicinity Jobs have emerged as a complement to traditional sources of labour demand data, such as the Job Vacancy and Wages Survey (JVWS). Ibrahim Abuallail, PhD Candidate, University of Ottawa, presented research relating to bias in OJPs and a proposed approach to effectively adjust OJP data to complement existing official data (such as from the JVWS) and improve the measurement of labour demand.
South Dakota State University degree offer diploma Transcriptynfqplhm
办理美国SDSU毕业证书制作南达科他州立大学假文凭定制Q微168899991做SDSU留信网教留服认证海牙认证改SDSU成绩单GPA做SDSU假学位证假文凭高仿毕业证GRE代考如何申请南达科他州立大学South Dakota State University degree offer diploma Transcript
Vicinity Jobs’ data includes more than three million 2023 OJPs and thousands of skills. Most skills appear in less than 0.02% of job postings, so most postings rely on a small subset of commonly used terms, like teamwork.
Laura Adkins-Hackett, Economist, LMIC, and Sukriti Trehan, Data Scientist, LMIC, presented their research exploring trends in the skills listed in OJPs to develop a deeper understanding of in-demand skills. This research project uses pointwise mutual information and other methods to extract more information about common skills from the relationships between skills, occupations and regions.
University of North Carolina at Charlotte degree offer diploma Transcripttscdzuip
办理美国UNCC毕业证书制作北卡大学夏洛特分校假文凭定制Q微168899991做UNCC留信网教留服认证海牙认证改UNCC成绩单GPA做UNCC假学位证假文凭高仿毕业证GRE代考如何申请北卡罗莱纳大学夏洛特分校University of North Carolina at Charlotte degree offer diploma Transcript
Solution Manual For Financial Accounting, 8th Canadian Edition 2024, by Libby...Donc Test
Solution Manual For Financial Accounting, 8th Canadian Edition 2024, by Libby, Hodge, Verified Chapters 1 - 13, Complete Newest Version Solution Manual For Financial Accounting, 8th Canadian Edition by Libby, Hodge, Verified Chapters 1 - 13, Complete Newest Version Solution Manual For Financial Accounting 8th Canadian Edition Pdf Chapters Download Stuvia Solution Manual For Financial Accounting 8th Canadian Edition Ebook Download Stuvia Solution Manual For Financial Accounting 8th Canadian Edition Pdf Solution Manual For Financial Accounting 8th Canadian Edition Pdf Download Stuvia Financial Accounting 8th Canadian Edition Pdf Chapters Download Stuvia Financial Accounting 8th Canadian Edition Ebook Download Stuvia Financial Accounting 8th Canadian Edition Pdf Financial Accounting 8th Canadian Edition Pdf Download Stuvia
STREETONOMICS: Exploring the Uncharted Territories of Informal Markets throug...sameer shah
Delve into the world of STREETONOMICS, where a team of 7 enthusiasts embarks on a journey to understand unorganized markets. By engaging with a coffee street vendor and crafting questionnaires, this project uncovers valuable insights into consumer behavior and market dynamics in informal settings."
Independent Study - College of Wooster Research (2023-2024) FDI, Culture, Glo...AntoniaOwensDetwiler
"Does Foreign Direct Investment Negatively Affect Preservation of Culture in the Global South? Case Studies in Thailand and Cambodia."
Do elements of globalization, such as Foreign Direct Investment (FDI), negatively affect the ability of countries in the Global South to preserve their culture? This research aims to answer this question by employing a cross-sectional comparative case study analysis utilizing methods of difference. Thailand and Cambodia are compared as they are in the same region and have a similar culture. The metric of difference between Thailand and Cambodia is their ability to preserve their culture. This ability is operationalized by their respective attitudes towards FDI; Thailand imposes stringent regulations and limitations on FDI while Cambodia does not hesitate to accept most FDI and imposes fewer limitations. The evidence from this study suggests that FDI from globally influential countries with high gross domestic products (GDPs) (e.g. China, U.S.) challenges the ability of countries with lower GDPs (e.g. Cambodia) to protect their culture. Furthermore, the ability, or lack thereof, of the receiving countries to protect their culture is amplified by the existence and implementation of restrictive FDI policies imposed by their governments.
My study abroad in Bali, Indonesia, inspired this research topic as I noticed how globalization is changing the culture of its people. I learned their language and way of life which helped me understand the beauty and importance of cultural preservation. I believe we could all benefit from learning new perspectives as they could help us ideate solutions to contemporary issues and empathize with others.
Falcon stands out as a top-tier P2P Invoice Discounting platform in India, bridging esteemed blue-chip companies and eager investors. Our goal is to transform the investment landscape in India by establishing a comprehensive destination for borrowers and investors with diverse profiles and needs, all while minimizing risk. What sets Falcon apart is the elimination of intermediaries such as commercial banks and depository institutions, allowing investors to enjoy higher yields.
2. Agenda
Basel II Overview
Operational Risk – Definition
Requirements of an Operational Risk Exposure Estimate
Loss Distribution Approach (LDA)
Segmenting Loss Data into Units of Measure
Literature on Frequency and Severity Modeling
Monte Carlo Simulation within an LDA based Operational Risk Exposure Model
Potential solutions for faster Monte Carlo Simulation
Description of the test environments utilized
Results from various methods of enhancement
2
3. Operational Risk Loss Events in the News
Barings Bank (1995) –$1.3 billion loss due to speculative trading performed by currency trader Nick Leeson. This
loss ultimately lead to the collapse of the bank.
Societe Generale (2008) - $7 billion loss based on the fraudulent activities of rogue futures trader Jerome Kerviel
DBS Bank, Ltd. (2010) - $310 million penalty imposed by the Monetary Authority of Singapore due to a seven hour
system-wide outage that left customers unable to use mobile, internet, and ATM services. Additionally, customers
were not able to make any debit or credit card transactions during the outage.
Citibank (2011) - $285 million settlement related to a failure to disclose to investors its role in the asset-selection
process for a hybrid Collateralized Debt Obligation the bank offered.
Multiple Banks (2012) - $25 billion in settlements and penalties regarding five large lenders’ improper foreclosure
practices between January 2008 and December 2011.
Note: Details for each of the events above were obtained from the Algorithmic’s FIRST database.
3
4. Basel II and Operational Risk
In December of 2007, the US Federal Reserve System finalized a document commonly referred to as the “Final Rules”
which set forth general requirements for the measurement of operational risk by large US financial institutions.1
These rules defined operational risk as the risk of loss resulting from inadequate or failed internal processes, people, and systems or
from external events (including legal risk but excluding strategic and reputational risk)
Seven Distinct Basel Loss Event Types2:
1. Internal Fraud
2. External Fraud 5. Damage to Physical Assets
3. Business Disruptions/System Failure 6. Clients, Products, and Business Practice Matters
4. Execution, Delivery and Process Management 7. Employee Practices and Workplace Safety Issues.
Other classifications are available to describe losses as defined by regulators and banks
E.g. - Business Lines, Regions of Operations, Causal of Loss Category, etc
The Final Rules require banks to estimate an operational risk exposure amount that corresponds to the 99.9th percentile of the
distribution of potential aggregate operational losses, as generated by the bank’s operational risk quantification system over a one-
year horizon.
Exposure estimates must:
a) Incorporate four data elements: Internal Loss Data, External Loss Data, Scenario Analysis Data, and Business
Environment/Internal Control Factor data.
b) Be calculated using systematic, transparent, verifiable, and credible methodologies
c) A separate exposure estimate must be calculated for each set of operational loss data demonstrating a statistically
distinct loss profile.
The banking industry has focused on the use of the Loss Distribution Approach (LDA) to calculate operational risk
exposure estimates.
1.Risk-Based Capital Standards: Advanced Capital Adequacy Framework – Basel II; Final Rule (2007), Federal Register 72(235), 69407 – 408. Also, see
Operational Risk – Supervisory Guidelines for the Advanced Measurement Approaches; BIS June 2011.
2. See the Appendix for examples of each loss event type
4
5. Overview of the Loss Distribution Approach (LDA)
Under the LDA, banks must segment their loss Segment Loss Data
data to obtain datasets that are not demonstrably into Homogenous
heterogeneous. Loss Datasets
These datasets are referred to as units of measure or
UOMs Loss Distribution Approach
These datasets are used for subsequent modeling
within the LDA
Frequency
The LDA models two primary components of Distribution
operational loss data: λ
Loss Frequency Aggregate Loss
# of loss events per year Distribution
The banking industry has widely accepted a
Poisson distribution as an appropriate distribution. Internal
and/or
Monte Carlo
Loss Severity Simulation
External
Fitting a parametric distribution to operational loss Loss
data is one of the biggest challenges in measuring Data
operational risk exposure.
Severity
Distribution Operational Risk Exposure
Monte Carlo Simulation is then utilized to compound the is estimated as the 99.9th
two distributions. percentile of the aggregate
loss distribution; a 1/1,000
A large number of simulations must be run to observe a sufficient
$ value of loss event year event *
number of losses to reasonably assess what a 1 in 1,000 year
event might look like…More on this shortly
* Banks typically sum the VaR estimates for their UOMs and
perform diversification modeling to move away from the
assumption of positive correlation.
5
6. Segmenting Loss Data into Units of Measure
Banks must segment their loss data to obtain datasets that are not demonstrably heterogeneous.
Which data classifications captured in the bank’s operational loss database best characterize the bank’s operational risk exposure?
Banks often capture a variety of details about individual loss events – E.g. Region of occurrence, Business Line, Basel Event Type
How granular should classification be?
Once an appropriate set of classifying variables has been identified, a natural starting point to narrow in on
homogenous datasets is to look at loss frequency and loss severity within the identified variables
Example Data:
Basel Loss Event Types In this example, the bank has
Loss Counts by Business
CPBP BDSF IF EPWS EF DPA EDPM Total
determined that 4 Business Lines
Line and Event Type
and the 7 Basel Loss Event Types are
Commercial Banking 20 50 - 20 550 50 - 690
in % of Total 2.90% 7.25% 0.00% 2.90% 79.71% 7.25% 0.00% 100.00% a reasonable representation of
Payment and Settlement
in % of Total
10
1.27%
30
3.80% 1.27%
10 15
1.90%
440 260
55.70% 32.91%
25 790
3.16% 100.00%
operational risk exposure.
Agency Services 80 130 10 30 850 10 5 1,115
in % of Total 7.17% 11.66% 0.90% 2.69% 76.23% 0.90% 0.45% 100.00%
Not every business line has a large
Other - 5 5 30 10 5 5 60 number of loss events.
in % of Total 0.00% 8.33% 8.33% 50.00% 16.67% 8.33% 8.33% 100.00%
Total 110 215 25 95 1,850 325 35 2,655 Within a business line, not every
in % of Total 4.14% 8.10% 0.94% 3.58% 69.68% 12.24% 1.32% 100.00% Basel Event Type classification level
Basel Loss Event Types has a large number of data
Loss Amounts by Business
Line and Event Type
(in $ MM)
CPBP BDSF IF EPWS EF DPA EDPM Total Basel Loss Event Type:
Commercial Banking $ 3.00 $ 18.00 $ 1.00 $ 10.00 $ 90.00 $ 1.00 $ - $ 123.00 BDSF and EF look distinct
in % of Total 2.44% 14.63% 0.81% 8.13% 73.17% 0.81% 0.00% 100.00%
Payment and Settlement $ 1.00 $ 7.00 $ 1.00 $ 4.00 $ 70.00 $ 25.00 $ 25.00 $ 133.00
in % of Total 0.75% 5.26% 0.75% 3.01% 52.63% 18.80% 18.80% 100.00% Business Line:
Agency Services $ 6.00 $ 240.00 $ 3.00 $ 2.00 $ 225.00 $ 5.00 $ 150.00 $ 631.00
in % of Total 0.95% 38.03% 0.48% 0.32% 35.66% 0.79% 23.77% 100.00% All 4 Business Lines might be distinct
Other $ - $ 3.00 $ 1.00 $ 15.00 $ 1.00 $ 3.00 $ 10.00 $ 33.00
in % of Total
Total $
0.00%
10.00 $
9.09%
268.00 $
3.03%
6.00 $
45.45%
31.00 $
3.03%
386.00 $
9.09%
34.00 $
30.30%
185.00 $
100.00%
920.00
Additional testing is required to
in % of Total 1.09% 29.13% 0.65% 3.37% 41.96% 3.70% 20.11% 100.00% identify homogenous datasets
Note: The data above are fiction, created for this example
6
7. Using R to Identify a Homogenous Loss Dataset
R can produce a variety of descriptive statistics, graphics, and hypothesis tests that are useful to evaluate whether
loss data should be merged (homogenous) or separated (heterogeneous).
Example: Is Business Disruptions & Systems Failure Loss Event Type statistically distinct from the
Commercial Banking Business Line?
Quantiles
Datasets Count Mean(Log) SD(Log) 50.0% 75.0% 90.0% 95.0% 98.0% 99.0% 99.5% Max
Commercial Banking 640 9.40 1.08 $ 7.89 $ 16.96 $ 65.39 $ 120.79 $ 281.52 $ 423.09 $ 777.20 $ 3,376.89
Business Disruptions & Systems Failure 215 11.55 2.08 $ 74.94 $ 424.44 $ 1,520.39 $ 5,904.35 $ 12,743.90 $ 17,255.26 $ 19,750.42 $ 19,954.35
- Quantiles in $ Thousands
Test Statistic pValue
Kolmogorov-Smirnov 0.53 0
Chi- Square 270.57 3.36674E-50
Anderson Darling 151.16 0
Conclusion:
The preponderance of evidence suggests that
Commercial Banking and BDSF are statistically distinct.
A separate risk exposure estimate should be calculated
for each of these datasets.
7
8. Using R to Identify Homogenous Loss Datasets
R offers the capability to produce a variety of descriptive statistics, graphics, and hypothesis tests that are useful to
evaluate whether loss data should be merged (homogenous) or separated (heterogeneous).
Example: Is Other Business Line statistically distinct from the Commercial Banking Business Line?
Quantiles
Datasets Count Mean(Log) SD(Log) 50.0% 75.0% 90.0% 95.0% 98.0% 99.0% 99.5% Max
Commercial Banking 640 9.40 1.08 $ 7.81 $ 17.61 $ 51.60 $ 106.22 $ 390.62 $ 617.59 $ 1,162.25 $ 3,142.99
Other 55 9.32 1.02 $ 7.70 $ 14.58 $ 40.38 $ 91.85 $ 138.79 $ 434.62 $ 607.40 $ 780.18
- Quantiles in $ Thousands
Test Statistic pValue
Kolmogorov-Smirnov 0.061 0.992
Chi- Square 6.580 0.884
Anderson Darling -0.998 0.627
Conclusion:
The preponderance of evidence suggests that we cannot
conclude the Commercial Banking and ‘Other’ Business
Lines are statistically distinct.
These data can be aggregated into a single data set for
frequency and severity modeling.
If a business rationale exists to keep these data sets
separate, banks may do so.
8
9. Frequency Distribution Fitting in R
Fitting a Frequency Distribution:
The banking industry has focused on the use of a Poisson distribution to model the frequency of operational loss
events.
The Poisson is parameterized by one parameter, λ, which is equivalent to the average frequency over the time horizon being
estimated (1 year).
Various methods are used to parameterize the Poisson distribution
Simple Annual Average
Bank identified internal and
Regression Analysis based on internal/external variables – See the function lm() in R external data characteristics might
Poisson Regression based on internal/external variables – See the function glm() in R help explain operational loss
frequency
Commercial Banking
Year Loss Counts
2005 76 Once a parameter estimate, λ ,has been identified,
2006 82 obtaining the density, distribution function, quantile
2007 94 function and random generation for the Poisson
2008 64
distribution is quite easy:
2009 90
2010 103 See dpois, ppois, rpois in R for more details.
2011 96
2012 85
Total 690
Average 86.25
9
10. Fitting a Severity Distribution
Fitting a Severity Distribution:
Many great authors have published overviews on the process for severity distribution fitting within the context of an
LDA model*.
The industry currently practices a variety of loss severity modeling techniques
Fitting a single parametric distribution to the entire dataset (e.g. – log normal, pareto, log gamma, weibull, etc.)
Fitting a mixture of parametric distributions to the loss severity data
Fitting multiple parametric distributions that have non-overlapping ranges (“Splicing”)
Extreme Value Theory (EVT) and the Peaks Over Thresholds Method
Challenges associated with fitting a severity distribution include:
1. The Final Rule asks banks to estimate a 1 in 1,000 year event based on less than 15 years of operational loss data
2. Data collection thresholds – Use of shifted distributions or truncated distributions?
3. Operational Loss Databases are often “living” – Loss severities, loss data classifications, and risk types can be modified
4. Data Paucity – In many cases banks have units of measure that have a small number of observations (< 1,000).
5. Undetected Heterogeneity of Datasets – Tests performed to identify heterogeneous datasets are not perfect at doing so
Small data sets can impede this effort.
6. Fat-Tailed Data – Banks are faced with UOMs that have a small number of observations which are often best described by a
heavily skewed distribution.
Limited data in the tail can result in volatile capital estimates (e.g. – capital can swing upwards or downwards by hundreds of
millions of $) based on the inclusion of a few events.
Volatile results can present subsequent challenges for obtaining senior management buy-in on risk exposure estimates.
* Please see the references slide at the end of this presentation for a short list of books and papers that provide additional detail on
operational risk modeling.
10
11. Fitting a Severity Distribution in R
Fitting a Severity Distribution:
A variety of optimization routines exist in R that are capable of fitting severity distribution to loss data.
Using the optim() in R, one needs to specify:
1. Density Function: - sum(densityFunction(x=data, log=TRUE))
2. Starting Parameters: Contingent upon the distribution being fit
3. Optimization Routine: Nelder-Mead, BFGS, SANN, etc.
See B. Bolker for more on optimization routines in R beyond the optim() function.
Fitting truncated severity distributions
The actuar package provides density, distribution, and quantile functions as well as random number generators for fat-tailed
distributions
See Nadarajah and Kotz for code that will facilitate the fitting of a truncated density, distribution, quantile function, and random
number generator.
Identifying a “best-fit” severity distribution to the loss data
QQ-Plot of the empircal data against the fitted distributions – plot(), qqplot()
Plot the empirical cdf against the fitted distribution – ecdf()
See truncgof R package and A. Chernobai, S. T. Rachev, F. J. Fabozzi for goodness-of-fit tests and some adjusted exploratory tools
that work with left truncated data.
Many packages exist that perform EVT severity distribution fitting:
See A. J. McNeil, R. Frey, P. Embrechts and the evir package in R.
Fitting and evaluating mixture distributions are more complex endeavors…
See the GAMLSS package in R and http://www.gamlss.org/
11
12. Overview of the Loss Distribution Approach (LDA)
Thus far we have discussed:
Segmentation of loss data to obtain datasets that are not Segment Loss Data
demonstrably heterogeneous. into Homogenous
Loss Frequency Modeling Loss Datasets
Loss Severity
Loss Distribution Approach
We have not yet discussed Monte Carlo Simulation…
Many simulations containing millions of iterations must be Frequency
run to observe a sufficient number of losses to reasonably Distribution
assess what a 1 in 1,000 year event might look like λ
This results in multiple days being lost to wait on code to Aggregate Loss
complete. # of loss events per year Distribution
Northern explored opportunities to parallelize Internal
Monte Carlo simulation with Revolution Analytics Monte Carlo
and/or Simulation
External
Loss
Data
Example Code:
# Randomly draw n frequency observations from a Poisson distribution,
then draw random severities from the specified truncated severity Severity
distribution, truncated at point a. Sum up each of the individual loss Distribution Operational Risk
amounts. Exposure is estimated as
f_tr <- function() { the 99.9th percentile of
sum(do.call("rtrunc", c(n=rpois(1, lambda), the aggregate loss
$ value of loss event distribution; a 1/1,000
spec=distName, a=a, parList)))
} year event
# Simulate a large number of iterations and replicate the simulation a
number of times to reduce sample noise
simuMatrix <- replicate(30, replicate(1e+6, f_tr()))
12
13. Monte Carlo Simulation Benchmarking Analysis
Northern Trust and Revolution Analytics Evaluate Various Methods to Enhance Monte Carlo Simulation
Use a different version of R: 32B, 64B (e.g. – Update your operating system)
Use various parallelization packages: doSNOW, doRSR, & doSMP, (doRSR & doSMP are Revolution Analytics product offerings)
Use multiple processors and/or machines:
Single node with multiple cores
Cluster of CPUs with multiple cores
Hardware Environments:
4-core laptop
3-node High Performing Cluster (HPC) on Amazon Cloud
Configured and run with 8-cores on each node
Each node was restricted from 16- to 8-cores
Metrics used to evaluate each method:
Elapsed Time by Step
Memory usage
13
14. Monte Carlo Benchmarking Highlights
Revolution Analytics’ parallelization can be easily scaled up from laptop/server to
the cluster using Revolution Analytics’ distributed computing capabilities
Parallelization greatly improves simulation performance
64bit is better
Elapsed time is linear in # of iterations
Performance improves with # of cores
Revo ~ Cran within a node (no MKL impact in this study)
doRSR slightly better than doSMP on a single server
64bit marginally better that 32bit
Performance scales with cluster resources
Memory use just driven by # of iterations
doRSR ~ doSMP Memory Trends
within a node Scales with # Cores
14
15. Take-Aways, Next Steps, and Contacts
Parallelizations Offers Business Enhancements:
Less time spent waiting on programs to complete
Means more time to analyze drivers of change (e.g. – underlying data changes)
More efficient management of computing resources
No need to manually manage/schedule programs
Scalability of the solution to available resources
Revolution Analytics’ parallelization routines are scalable to the resources available
Contact Information:
Dave Humke, Northern Trust, Vice President, (dh98@ntrs.com)
Derek Norton, Revolution Analytics, (derek.norton@revolutionanalytics.com)
15
16. Appendix – Basel Loss Event Type Definition
Event Type Category Definition Categories (Level 2) Activity Examples (Level 3)
(Level 1)
Internal Fraud Loss due to acts of a type Unauthorized Transactions not reported (intentional)
intended to defraud, Activity Transaction type unauthorized (with monetary loss)
misappropriate property or Mismarking of position (intentional)
circumvent regulations, the
Theft and Fraud Fraud / credit fraud / worthless deposits
law or company policy,
excluding diversity / Theft / extortion / embezzlement / robbery
discrimination events, which Misappropriation of assets
involves at least one internal Forgery
party. Check kiting
Smuggling
Account take-over / impersonation, etc.
Tax non-compliance / evasion (willful)
Bribes / kickbacks
Insider trading (not on firm's account)
External Fraud Losses due to acts of a type Theft and Fraud Theft / robbery
intended to defraud, Forgery
misappropriate property or Check kiting
circumvent the law, by a
Systems Security Hacking damage
third party
Theft of information (with monetary loss)
Employment Losses arising from acts Employee Relations Compensation, benefit, termination issues
Practices and inconsistent with Organized labor activities
Workplace Safety employment, health or safety Safe Environment General liability (slips and falls, etc.)
laws or agreements, from
Employee health & safety rules and events
payment of personal injury
claims, or from diversity / Workers compensation
discrimination events. Diversity & All discrimination types
Discrimination
16
17. Appendix – Basel Loss Event Type Definitions (Continued)
Event Type Category Definition Categories (Level 2) Activity Examples (Level 3)
(Level 1)
Clients, Products & Losses arising from an Suitability, Fiduciary breaches / guideline violations
Business Practice unintentional or negligent Disclosure & Suitability / disclosure issues (KYC, etc.)
failure to meet a professional Fiduciary Retail consumer disclosure violations
obligation to specific clients
Breach of privacy
(including fiduciary and
suitability requirements), or Aggressive sales
from the nature or design of Account churning
a product. Misuse of confidential information
Lender liability
Improper Business or Antitrust
Market Practices Improper trade / market practice
Market manipulation
Insider trading (on firm's account)
Unlicensed activity
Money laundering
Product Flaws Product defects (unauthorized, etc.)
Model errors
Selection, Failure t investigate client per guidelines
Sponsorship & Exceeding client exposure limits
E
Advisory Activities Disputes over performance or advisory activities
Damage to Physical Losses arising from loss or Disasters and Other Natural disaster losses
Assets damage to physical assets Events Human losses from external sources (terrorism,
from natural disaster or other vandalism)
Business Disruption Losses arising from disruption Systems Hardware
& Systems Failures of business or system Software
failures Telecommunications
Utility outage / disruptions
17
18. Appendix – Basel Loss Event Type Definitions (Continued)
Event Type Category Definition Categories (Level 2) Activity Examples (Level 3)
(Level 1)
Execution, Delivery Losses from failed Transaction Miscommunication
& Process transaction processing or Capture, Execution Data entry, maintenance or loading error
Management process management, from & Maintenance Missed deadline or responsibility
relations with trade
Model / system misoperation
counterparties and vendors
Accounting error / entity attribution error
Other task misperformance
Delivery failure
Collateral management failure
Reference data maintenance
Monitoring & Failed mandatory reporting obligation
Reporting Inaccurate external report (loss incurred)
Customer Intake & Client permissions / disclaimers missed
Documentation Legal documents missing / incomplete
Customer / Client Unapproved access given to accounts
Account Incorrect client records (loss incurred)
Management Negligent loss or damage of client assets
Trade Non-client counterparty misperformance
Counterparties Misc. non-client counterparty disputes
Vendors & Suppliers Outsourcing
Vendor disputes
18
19. Appendix - References
References on Loss Distribution Approach Modeling, Frequency and Severity Fitting, and Monte Carlo Simulation:
1. A. Chernobai, S. T. Rachev, F. J. Fabozzi (2005), Composite Goodness-of-Fit Tests for Left-Truncated Samples, Technical report,
University of California Santa Barbara
2. A. J. McNeil, R. Frey, P. Embrechts (2005), Quantitative Risk Management: Concepts, Techniques, and Tools, Princeton University
Press, Princeton
3. B. Bolker (2007), Optimization and All That, Draft of Chapter 7 of B. Bolker (2008), Ecological Models and Data in R, Princeton
University Press, Princeton
4. G.J. McLachlan, D. Peel (2000), Finite Mixture Models, Wiley & Sons, New York
5. H. Panjer (2006), Operational Risk: Modeling Analytics, Wiley & Sons, New York, p. 293.
6. K. Dutta, J. Perry (2006), A Tale of Tails: An Empirical Analysis of Loss Distribution Models for Estimating Operational Risk Capital,
Working Paper No. 06-13, Federal Reserve Bank of Boston.
7. M. Moscadelli (2004), The Modelling of Operational Risk: Experience with the Analysis of the Data Collected by the Basel Committee,
Temi di Discussione No. 517, Banca d’Italia.
8. P. de Fontnouvelle, E. Rosengren, J. Jordan (2007), Implications of Alternative Operational Risk Modeling Techniques, In: M. Carey
and R.M. Stulz (eds), The Risks of Financial Institutions, University of Chicago Press, pp. 475-512.
9. S.A. Klugman, H.H. Panjer, G.E. Willmot (2008), Loss Models: From Data to Decisions, 3rd ed., Wiley & Sons, Hoboken, NJ
10. S. Nadarajah, S. Kotz (2006), R Programs for Computing Truncated Distributions, Journal of Statistical Software 16(2)
19