| |
 |
| Paper: |
Improving the Statistical Methodology of Astronomical Data Analysis |
| Volume: |
25, Astronomical Data Analysis Software and Systems I |
| Page: |
237 |
| Authors: |
Feigelson, Eric D.; Babu, Gutti Jogesh |
| Abstract: |
Contemporary observational astronomers are generally unfamiliar with the extensive advances made in mathematical and applied statistics during the past several decades. Astronomical problems can often be addressed by methods developed in statistical fields such as spatial point processes, density estimation, Bayesian statistics, and sampling theory. The common problem of bivariate linear regression illustrates the need for sophisticated methods. Astronomical problems often require combinations of ordinary least-squares lines, double-weighted and errors-in-variables models, censored and truncated regressions, each with its own error analysis procedure. The recent conference Statistical Challenges in Modern Astronomy highlighted issues of mutual interest to statisticians and astronomers including clustering of point processes and time series analysis. We conclude with advice on how the astronomical community can advance its statistical methodology with improvements in education of astrophysicists, collaboration and consultation with professional statisticians, and acquisition of new software. |
|
|
|
|
|
