FAQ www.rmetrics.org
2004-10-11 Rmetrics 200.10058
What is Rmetrics?
Rmetrics is a collection of several hundreds of functions which may be useful for teaching "Financial Engineering" and "Computational Finance". These functions are available for R, GNUs S. This is a freely available language and environment for statistical computing and graphics which provides a wide variety of statistical and graphical techniques.
The functions have their source in algorithms and functions written by myself, my students, and many other authors. A major aim is to bring financial algorithms and concepts together under a common software platform and to make it public available mainly for teaching financial engineering and computational finance. Rmetrics is not part of CRAN, Rmetrics is an initiative by its own. Rmetrics has some aims similar like Bioconductor. Rmetrics is an open source and open development software project.
What are the roots of Rmetrics?
The basic R port from which Rmetrics originated was already initiated in 1999 as an outcome of lectures held by Diethelm Würtz on topics in econophysics at ETH Zürich. Meanwhile, the family of the Rmetrics packages includes four members dealing with the following subjects: fBasics - Markets, Basic Statistics, Date and Time, fSeries - The Dynamical Process Behind Financial Markets, fExtremes - Beyond the Sample, Dealing with Extreme Values, and fOptions - The Valuation of Options. Two other packages are under current development, fBonds and fPortfolio.
What is included in the fBasics Package?
The package fBasics covers the management of economic and financial market data. Included are functions to download economic indicators and financial market data from the Internet. Distribution functions relevant in finance are added like the asymmetric stable, the hyperbolic and the inverse normal gaussian distribution function to compute densities, probabilities, quantiles and random deviates. Estimators to fit the distributional parameters are also available. Some additional hypothesis tests for the investigation of correlations, dependencies and other stylized facts of financial time series can also be found in this package. Furthermore, for date and time management a holiday database for all ecclestial and public holidays in the G7 countries and Switzerland is provided together with a database of daylight saving times for financial centers around the world. Special calendar management functions were implemented to create easily business calendars for exchanges. A collection of functions for filtering and outlier detection of high frequency foreign exchange data records collected from Reuters' data feed can also be found together with functions for de-volatilization and de-seasonalization of the data.
What is included in the fSeries Package?
The package fSeries covers topics from the field of financial time series analysis including ARIMA, GARCH, Regression, and Feedforward Neural Network modelling. This library tries to bring together the content of existing R-packages with additional new functionality on a common platform. The collection comes with functions for testing various aspects of financial time series, including unit roots, independence, normality of the distribution, trend stationary, co-integration and neglected non-linearities. Furthermore functions for testing for higher serial correlations, for heteroskedasticity, for autocorrelations of disturbances, for linearity, and functional relations are also provided. Technical analysis and benchmarking is another major issue of this package. The collection offers a set of the most common technical trading indicators together with functions for charting and benchmark measurements. For building trading models functions for a rolling market analysis are available. A new additional chapter on modeling long memory behavior including moment methods, periodgram analysis, whittle estimator, and wavelet analysis is planned to be added in the near future.
What is included in the fExtremes Package?
The package fExtremes covers topics from the field what is known as extreme value theory. The package has functions for the exploratory data analysis of extreme values in insurance, economics, and finance applications. Included are plot functions for empirical distributions, quantile plots, graphs exploring the properties of exceedences over a threshold, plots for mean/sum ratio and for the development of records. Furthermore functions for preprocessing data for extreme value analysis are available offering tools to separate data beyond a threshold value, to compute blockwise data like block maxima, and to de-cluster point process data. One major aspect of this package is to bring together the content of already existing R-packages with additional new functionality for financial engineers on a common platform investigating fluctuations of maxima, extremes via point processes, and the extremal index. A new additional chapter on risk measures, stress testing and copulae is planned to be added in the near future.
What is included in the fOptions Package?
The package fOptions covers the valuation of options including topics like the basics of option pricing in the framework of Black and Scholes, including almost 100 functions for exotic options pricing, including the Heston-Nandi option pricing approach mastering stochastic volatility, and Monte Carlo simulations together with generators for low discrepancy sequences. Beside the Black and Scholes option pricing formulas, functions to valuate other plain vanilla options on commodities and futures, and function to approximate American options are also available. Some binomial tree models are implemented. The exotic options part comes with a large number of functions to valuate multiple exercise options, multiple asset options, lookback options, barrier options, binary options, Asian options, and currency translated options. Some functions for the investigation of exponential Brownian motion in the context of Asian option valuation have been recently added.
Is Rmetrics Open Source Software?
Rmetrics has a commitment to full open source code development and distribution. All contributions included in the Rmetrics packages are expected to exist under an open source license such as GPL2. The reasons for this commitment are the ability to test, to extend and to improve the software in a convenient way, to encourage excellent scientific computing and statistical practice in financial engineering and computational finance, and to provide a workbench of tools that allow to explore and expand the methods used to analyze financial market data and to valuate financial instruments.
What Platforms are supported by Rmetrics?
Rmetrics is primarily build and maintained under MS Windows XP.
The latest source packages are located in the source directory on the Rmetrics Server, and the latest binary files for the Windows OS are located in the download directory themselves. The DESCRIPTION files hold the most recent version number, please check. The source code of the productive packages can be downloaded from the CRAN Server, also the binary packages for Windows, Mac OSX and Linux operated computers. Rmetrics is also availalble in form of Debian Packages and part of the latest Knoppix Quantian CD.
Why did you start the Rmetrics Port for Microsoft Software?
In the financial community and also in my lectures Windows is the mostly used operating system. To teach financial engineers it became quite natural for me to work under Windows. For a broad distribution and acceptance of Rmetrics I decided to devellop the software under Windows 2000/XP. That is business reality, and it was a good decision.
There are no many people behind Rmetrics, currently it's only me. I have my job as a lecturer at the Institute of Theoretical Physics at ETH Zurich, and I'm the senior partner of an ETH spin-off software company, Finance Online. So I have several responsibilities to take, and as a consequence things might go slow ... The growing Rmetrics collection is based on many statistical and financial functions which were contributed by myself, my students, or were ported from many other sources during the last seven years since I started my lectures in "Econophysics" at ETH. I'm aware that the work is by far not complete. Parts of the software are still untested, and may contain some bugs. Contributors are welcome!
Diethelm Würtz