Redshift measurements are vital to modern astrophysics. Using spectral line shifts is an accurate method to calculate redshift for a distant galaxy but the observation time to develop the spectrum is quite high and the result's accuracy has a high dependence on SNR, which drastically limits its use to applications which require us to find redshifts for a large number of objects.

A faster, though less accurate way is to correlate photometric data with redshifts. We used certain machine learning techniques to estimate photometric redshifts.

The full report can be downloaded here.