SIMBAD references

In our former paper I, we showed on the Sun that different active regions possess unique intensity profiles on the Ca II H & K lines. We now extend the analysis by showing how those properties can be used on real stellar observations, delivering more powerful activity proxies for radial velocity correction. More information can be extracted on rotational time-scale from the Ca II H & K lines than the classical indicators: S-index and $\log (R^{^{\prime }}_{\text{HK}})$. For high-resolution HARPS observations of $\alpha$ Cen B, we apply a principal and independent component analysis on the Ca II H & K spectra time-series to disentangle the different sources that contribute to the disc-integrated line profiles. While the first component can be understood as a denoised version of the Mount-Wilson S-index, the second component appears as powerful activity proxies to correct the radial velocities induced by the inhibition of the convective blueshift in stellar active regions. However, we failed to interpret the extracted component into a physical framework. We conclude that a more complex kernel or bandpass than the classical triangular of the Mount Wilson convention should be used to extract activity proxies. To this regard, we provide the first principal component activity profile obtained across the spectral type sequence between M1V to F9V type stars.