Massive lossless data compression and multiple parameter estimation from galaxy spectra

• DOI: 10.1046/j.1365-8711.2000.03692.x
• Publication Date: 2003-03-12T13:15:33Z
• Authors: A. F. Heavens; R. Jimenez; O. Lahav
• Keywords: Data set

We present a method for radical linear compression of data sets where the are dependent on some number M parameters. show that, if noise in is independent parameters, we can form combinations which contain as much information about all parameters entire set, sense that Fisher matrices identical; i.e. lossless. explore how these compressed numbers fare when and method, though not precisely lossless, increases errors by very modest factor. The general, but illustrate it with problem well-suited: galaxy spectra, typically consist ∼103 fluxes, properties set handful such age, parametrized star formation history. spectra reduced to small data, connected physical processes entering problem. This offers possibility large increase speed determining an important consideration reach 106 size, complexity model increases. In addition this practical advantage, may offer classification scheme based rather directly processes.