Density Estimation Trees as fast non-parametric modelling tools

17th International Workshop on Advanced Computing and Analysis Techniques in Physics Research

物理学;计算机科学

Anderlini, Lucio^1

Istituto Nazionale di Fisica Nucleare, Sezione di Firenze, via Sansone 1, Sesto Fiorentino

50019, Italy^1

关键词: Density estimation;    Detector calibration;    Fast simulation;    Non-parametric modelling;   

Others  :  https://iopscience.iop.org/article/10.1088/1742-6596/762/1/012042/pdf
DOI  :  10.1088/1742-6596/762/1/012042

学科分类：计算机科学（综合）

来源: IOP

【 摘 要 】

A Density Estimation Tree (DET) is a decision trees trained on a multivariate dataset to estimate the underlying probability density function. While not competitive with kernel techniques in terms of accuracy, DETs are incredibly fast, embarrassingly parallel and relatively small when stored to disk. These properties make DETs appealing in the resource-expensive horizon of the LHC data analysis. Possible applications may include selection optimization, fast simulation and fast detector calibration. In this contribution I describe the algorithm and its implementation made available to the HEP community as a RooFit object. A set of applications under discussion within the LHCb Collaboration are also briefly illustrated.