Prior to 1947, the activity of the Sun was assessed by the relative numbers of sunspots (W). The 10.7 cm radio emission (frequency of 2.8 GHz) for observations of the variability of radiation of chromosphere and the lower corona (F10.7) became used f
rom 1947. For the F10,7 are available more detailed observational archive data, so this activity index more often than the other indices is used in the prediction and monitoring of the solar activity. We have made the analysis of time series of F10.7 with the use of different mother wavelets: Daubechies 10, Symlet 8, Meyer, Gauss 8 and Morlet. Wavelet spectrum allows us not only to identify cycles, but analyze their change in time. Each wavelet has its own characteristic features, so sometimes with the help of different wavelets it can be better identify and highlight the different properties of the analyzed signal. We intended to choose the mother wavelet, which is more fully gives information about the analyzed index F10.7. We have received, that all these wavelets show similar values to the maximums of the cyclic activity. However, we can see the difference when using different wavelets. There are also a number of periods, which, perhaps, are the harmonics of main period. The mean value of 11-year cycle is about 10.2 years. All the above examples show that the best results we get when using wavelets Morley, Gauss (real-valued) and multiparameter family of wavelets Morley and Gauss (complex-valued).
The correlation coefficients of the linear regression of six solar indices versus F10,7 were analyzed in solar cycles 21, 22 and 23. We also analyzed the interconnection between these indices and F10,7 with help of the approximation by the polynomial
s of second order. The indices weve studied in this paper are: Wolf numbers - W, 530,3 nm coronal line flux - F530, the total solar irradiance - TSI, Mg II UV-index 280 nm core-to-wing ratio, Flare Index - FI and Counts of flares. In the most cases the regressions of these solar indices versus F10,7 are close to linear except the moments of time near to the minimums and maximums of 11-year activity. For the linear regressions we found that the values of correlation coefficients Kcorr(t) for the indices versus F10,7 and W show the cyclic variations with periods approximately equal to the to half length of 11-year cycle - 5,5 years approximately.
