Speaker:Dr Richard Nickl (Statslab
Venue: Winstanley Lecture Theatre
Time: 07/03/2011 20:30, drinks from 20:15
Even after C.F. Gauss had published the Disquistiones Arithmeticae, that contain some of his most fundamental number-theoretical work, in 1801, he was known only to specialists in Europe. This changed drastically after he predicted, late in 1801, the position of the planet Ceres after it had ‘disappeared behind the sun’. Gauss immediately achieved fame throughout Europe. A technique that Gauss used in this prediction, as well as in much of his subsequent observational work, was the method of least squares to correct ‘intrinsic’ measurement errors in an efficient way. This method, in conjunction with the normal or ‘Gaussian’ distribution, has been a cornerstone of modern statistics ever since, reaching into the foundations of likelihood-based inference. I shall discuss the main mathematical and methodological ideas behind Gauss’ invention and trace some key aspects of its history into the 21st century, highlighting a variety of key statistical ideas that derive from it.