My major research interests are in the field of particle physics.
<I am working on Fermilab experiment, BooNE (Booster Neutrino Experiment) which is searching for a small neutrino mass by looking for oscillations of muon neutrinos into electron neutrinos. The Los Alamos Neutrino Detector (LSND) reported a possible signal, and our experiment has made a definitive test. We do not observe the LSND signal if it is due to standard oscillations between two states. It is still possible that the LSND results are due to exotic interactions and we are studying a low energy anomaly that, possibly, may be related to this. I supplied some monitoring counters for the experiment, contributed to the neutrino flux program, wrote one of three event reconstruction programs, and introduced the new technique of boosted decision trees into MiniBooNE (and physics generally). This technique was used as the basis for one of our two oscillation analyses. We are now running an anti-neutrino beam and will search for oscillations and our low energy anomaly there. In addition there are a number of quite useful cross section measurements that we are making. One of these, an analysis of charged current quasi-elastic events in our energy region is now in press. It has two orders of magnitude more events than any previous experiment.
Postdoc collaborators: Hai-jun Yang
Page of gif and other picture files: Gif and other files
This is a pdf file of the Middendorf thesis, which uses the boosting technique for biological investigations: Middendorf thesis
This is the ppt file of my boosting talk in early 2005. (There are a bunch of miscellaneous and outdated slides after the end--Please ignore them): Boosting talk 2005
The next file is the gzipped tar file for the C++ version of my boost program. The latest modification is May 8, 2007 and contains important improvements and bug fixes, including an error that prevented softscore from working on data. If you copy this file, please e-mail me at byronroe@umich.edu so that I can notify you of important changes. This is the tar file for the C++ version of m-boost.
The next file is the gzipped tar file for the FORTRAN version of my boost program. The latest modification is May 8, 2007 and fixes a bug in the data_loop_over_trees routine. The February 17, 2007 contains important improvements and bug fixes, including an error that prevented softscore from working on data. If you copy this file, please e-mail me at byronroe@umich.edu so that I can notify you of important changes. This is the tar file for the FORTRAN version of m-boost
This link is for a file of events and background for the Banff meeting (2006). The file is a formatted file. In the first line is the the number of signal events followed by the number of background events. The signal events come first, followed by the background events. Each line, after the first line has the 50 particle ID variables for one event. 50 variable data file