This article pretends to show how can be removed the noise and artifacts from electrocardiographic signal, such as Principal component analysis and Principal component analysis, they will be applied on a data base of electrocardiogram who have been taken on humans been. The FastICA algorithm using negentropy as cost function, shows a better result over its variant call JADE, in turn FastICA algorithm shows superiority in noise elimination in comparison of a classical filtering scheme composed for a FIR filter bank