|
|
||
   School of Computational Sciences Home Students Research Simulation gallery Publications Teaching (Courses) Data archive Annual conference Computing resources ├ OMEGA ├ WRF └ BOOT Related links About our webpage Contact us |
 |
BOOTCAMP has developed a software package called BOOT that implements a set of statistical performance measures for evaluating dispersion and other numerical models. The BOOT software originally included the following statistical performance measures: the fractional bias (FB), geometric mean bias (MG), normalized mean square error (NMSE), geometric mean variance (VG), correlation coefficient (R) and fraction of predictions within a factor of two of observations (FAC2). BOOT can be easily updated to include other performance measures. For example, the software was recently extended to also include the procedures adopted by the American Society for Testing and Materials (ASTM) Guide 6589-00 (ASTM 2000) and the Measure of Effectiveness (MOE) suggested by the Institute for Defense Analyses (IDA). BOOT has been extensively used by other scientists for model performance evaluation. The culmination of these evaluation studies led to a recent recommendation of the acceptance criteria for dispersion models (Chang and Hanna 2004).In addition to calculating quantitative performance measures using BOOT, it is also essential to conduct exploratory data analysis to visualize model performance using different types of plots. Some of the commonly-used plots include scatter plots, quantile-quantile plots, residual box plots, conditional scatter plots, VG/MG plots and polar plots. See the
contact page for detailed
information on
contacting us. |
