The characteristics of the radwastes generated from a nuclear power plants cannot be directly described due to the various problems associated with it. The scaling factors which correlate the hard-to-measure nuclides to the easy-to-measure nuclides are proposed and used as a tool describing indirectly the characteristics of the radwastes. However, the scaling factors are generic only for selective nuclides and must be calculated repeatedly whenever the plant condition changes substantially. To solve the problems, the characteristics of the radwastes must be correlated to the plant``s operational history. Since the existing sample data could not provide the information on the characteristics of the radwastes according to the plant``s operational history, the characteristics of the radwastes should be correlated to the other characteristics of radwaste which can not only describe roughly the plant condition, but also can be easily obtained in plant routine basis. The two variables, independent and dependent, which are correlated in this study are adjusted in order that causal relationship between the two variables can be easily identified, and are all assumed to have the lognormal distribution. By graphical analysis, the independent variable is determined as $Q(Co^{60})+Q(Cs^{137})$ and the dependent variable as $Q(i)/[Q(Co^{60})+Q(Cs^{137})]^2$, where Q means the activity concentration and imeans the hard-to-measure nuclide. The correlation relation of which log mean dispersion is less than 10 is presented as calculation formula. The correlation relation for $C^{14},\; Sr^{90},\; I^{129}$(they are very difficult to characterize previously) and $Fe^{55}$ turn out to be very good for all types of waste. Since the formation and disappearance mechanism of TRU is more complex than that of corrosion and fission products, the correlation relations for TRU have mostly log mean dispersions which are greater than 10. The quantity which the characteristics of TRU are corre...