Search

RooStats is a project to create advanced statistical tools required for the analysis of LHC data, with emphasis on discoveries, confidence intervals, and combined measurements. The idea is to provide the major statistical techniques as a set of C++ classes with coherent interfaces, which can be used on arbitrary model and datasets in a common way. The classes are built on top of RooFit, which provides a very convenient functionality for modeling the probability density functions or the likelihood functions, required as inputs for any statistical technique. Furthermore, RooFit provides via the RooWorkspace class, the functionality for easily creating models, for analysis combination and for digital publication of the likelihood function and the data. We will present in detail the design and the implementation of the different statistical methods of RooStats. These include various classes for interval estimation and for hypothesis test depending on different statistical techniques such as those based on the likelihood function, or on frequentists or bayesian statistics. These methods can be applied in complex problems, including cases with multi parameter of interests and various nuisance parameters. We will also show some example of usage and we will describe the results and the statistical plots obtained by running the RooStats methods.

The Parallel ROOT Facility - PROOF is a distributed analysis system which allows to exploit inherent event level parallelism of high energy physics data. PROOF can be configured to work with centralized storage systems, but it is especially effective together with distributed local storage systems - like Xrootd, when data are distributed over computing nodes. It works efficiently on different types of hardware and scales well from a multi-core laptop to large computing farms. From that point of view it is well suited for both large central analysis facilities and Tier 3 type analysis farms. PROOF can be used in interactive or batch like regimes. The interactive regime allows user to work with typically distributed data from ROOT command prompt and get a real time feedback on analysis progress and intermediate results. We will discuss our experience with PROOF in the context of ATLAS Collaboration distributed analysis. In particular we will discuss PROOF performance in various analysis scenarios and in multi-user, multi-session environment. We will also describe PROOF integration with ATLAS distributed data management system and prospects of running PROOF on geographically distributed analysis farms.

-expect new roofit/roostats package from Kyle Cranmer and Wouter Verkerke -material in branches must be moved before middle of next week -developments by Ilka/Roj will be introduced after the release -fixes to get "make static" and the code checker working again -THtml must be modified to support the new dir structure (urgent)

In anticipation of data taking, ATLAS has undertaken a program of work to develop an explicit state representation of the experiment's complex transient event data model. This effort has provided both an opportunity to consider explicitly the structure, organization, and content of the ATLAS persistent event store before writing tens of petabytes of data (replacing simple streaming, which uses the persistent store as a core dump of transient memory), and a locus for support of event data model evolution, including significant refactoring, beyond the automatic schema evolution capabilities of underlying persistence technologies. ATLAS has encountered the need for such non-trivial schema evolution on several occasions already. This paper describes the state representation strategy (transient/persistent separation) and its implementation, including both the payoffs that ATLAS has seen (significant and sometimes surpising space and performance improvements, the extra layer notwithstanding, and extremely general schema evolution support) and the costs (additional and relatively pervasive additional infrastructure development and maintenance). The paper further discusses how these costs are mitigated, and how ATLAS is able to implement this strategy without losing the ability to take advantage of the (improving!) automatic schema evolution capabilities of underlying technology layers when appropriate. Implications of state representations for direct ROOT browability, and current strategies for associating physics analysis views with such state representations, are also described.

As we near the collection of the first data from the Large Hadron Collider, the ATLAS collaboration is preparing the software and computing infrastructure to allow quick analysis of the first data and support of the long-term steady-state ATLAS physics program. As part of this effort considerable attention has been payed to the "Analysis Model", a vision of the interplay of the software design, computing constraints, and various physics requirements. An important input to this activity has been the experience of Tevatron and B-Factory experiments, one topic which was explored discussed in the ATLAS October 2006 Analysis Model workshop. Recently, much of the Analysis Model has focused on ensuring the ATLAS software framework supports the required manipulations of event data; the event data design and content is consistent with foreseen calibration and physics analysis tasks; the event data is optimized in size, access speed, and is accessible both inside and outside the software framework; and that the analysis software may be developed collaboratively.

The EventView Analysis Framework is currently the basis for much of the analysis software employed by various ATLAS physics groups (for example the Top, SUSY, Higgs, and Exotics working groups). In ATLAS's central data preparation, this framework provides an assessment of data quality and the first analysis of physics data for the whole collaboration. An EventView is a self-consistent interpretation of a physics event or equivalently the state of a specific analysis. Analyses are constructed at runtime by chaining and configuring modular components consisting of tools, C++ implementation of specific analysis algorithms, and modules, python grouping and configuration of various tool. A large common library of general tools and modules serve as the building blocks of nearly all of the steps of any analysis. The output is multiple simultaneous EventViews of every event, typically reflecting different choices of selections, reconstruction algorithms, combinatoric assignments, or input data (eg full or fast reconstruction or truth).

The LHC has enormous discovery potential and poses many new statistical challenges. We will review the requirements for statistical toolkit for the LHC experiments, including the ability to combine the results of multiple measurements, incorporate systematic uncertainty, and facilitate the technical aspects of sharing code. In recent years, several statistical methods have been proposed to incorporate systematic uncertainty ranging from Bayesian, to fully Frequentist, to hybrid techniques. It will be shown that these various methods can have quite different properties, which makes it imperative that the toolkit allows for one to simultaneously evaluate the same problem with multiple techniques. With these considerations in mind, we set out to develop a statistical toolkit for ROOT. The RooFit package has a large user community and an abstract PDF class that can support both Bayesian and Frequentist interpretations; thus, we decided to base the statistical toolkit on the existing RooFit components and call the package RooStats. Finally, we will give some examples and outline our plans for further development.