Estimation of Full Sky Power Spectrum between Intermediate to Large Angular Scales from Partial Sky CMB Anisotropies using Artificial Neural Network

Abstract

Reliable extraction of cosmological information from observed cosmic microwave background (CMB) maps requires accurate removal of foreground components. Since some regions of the sky are inevitably contaminated by strong foreground signals, it may be desired to excise such strongly contaminated regions even after a foreground subtraction has been performed. In this article, we employ an artificial neural network (ANN) to predict the full sky CMB angular power spectrum from the partial sky spectrum obtained from masked CMB temperature anisotropy map, at a high pixel resolution. We use a simple ANN architecture with one hidden layer containing 895 neurons. Using 1.2 × 10 5 training samples of full sky and corresponding partial sky CMB angular power spectra at Healpix pixel resolution parameter N side = 256, we show that predicted spectrum by our ANN agrees well with the target spectrum at each realization for the multipole range 2 ≤ l ≤ 512. The predicted spectra are statistically unbiased and they preserve the cosmic variance accurately. Statistically, the differences between the mean predicted and underlying theoretical spectra are within 3σ. Moreover, the probability densities obtained from predicted angular power spectra agree very well with those obtained from 'actual' full sky CMB angular power spectra for each multipole. Equally interesting is that the correlation matrix calculated from predicted angular power spectra does not show any signature of correlation between the angular power spectrum of different multipoles. Our work shows that the predicted spectra do not need to