arXiv subjects

I. Narsky

Publications and source records attributed to I. Narsky.

3 recordsLinked to original sources

arXiv2005

Optimization of Signal Significance by Bagging Decision Trees

An algorithm for optimization of signal significance or any other classification figure of merit suited for analysis of high energy physics (HEP) data is described. This algorithm trains decision trees on many bootstrap replicas of training data with each tree required to optimize the signal significance or any other chosen figure of merit. New data are then classified by a simple majority vote of the built trees. The performance of this algorithm has been studied using a search for the radiative leptonic decay B->gamma l nu at BaBar and shown to be superior to that of all other attempted classifiers including such powerful methods as boosted decision trees. In the B->gamma e nu channel, the described algorithm increases the expected signal significance from 2.4 sigma obtained by an original method designed for the B->gamma l nu analysis to 3.0 sigma.

physics.data-an↗

arXiv2005

StatPatternRecognition: A C++ Package for Statistical Analysis of High Energy Physics Data

Modern analysis of high energy physics (HEP) data needs advanced statistical tools to separate signal from background. A C++ package has been implemented to provide such tools for the HEP community. The package includes linear and quadratic discriminant analysis, decision trees, bump hunting (PRIM), boosting (AdaBoost), bagging and random forest algorithms, and interfaces to the standard backpropagation neural net and radial basis function neural net implemented in the Stuttgart Neural Network Simulator. Supplemental tools such as bootstrap, estimation of data moments, and a test of zero correlation between two variables with a joint elliptical distribution are also provided. The package offers a convenient set of tools for imposing requirements on input data and displaying output. Integrated in the BaBar computing environment, the package maintains a minimal set of external dependencies and therefore can be easily adapted to any other environment. It has been tested on many idealistic and realistic examples.

physics.data-an↗

arXiv2003

The BaBar Event Building and Level-3 Trigger Farm Upgrade

The BaBar experiment is the particle detector at the PEP-II B-factory facility at the Stanford Linear Accelerator Center. During the summer shutdown 2002 the BaBar Event Building and Level-3 trigger farm were upgraded from 60 Sun Ultra-5 machines and 100MBit/s Ethernet to 50 Dual-CPU 1.4GHz Pentium-III systems with Gigabit Ethernet. Combined with an upgrade to Gigabit Ethernet on the source side and a major feature extraction software speedup, this pushes the performance of the BaBar event builder and L3 filter to 5.5kHz at current background levels, almost three times the original design rate of 2kHz. For our specific application the new farm provides 8.5 times the CPU power of the old system.

physics.ins-det↗