About predicting (estimating) the effect of the temperature on the thermal conductivity of alloys, please check my answer given elsewhere at this forum: https://www.researchgate.net/post/How_can_I_find_temperature_dependent_properties_of_a_materialfor_example_steel
For the effect of the composition on the thermal conductivity of alloys, you may check: https://www.researchgate.net/post/Is_there_a_model_for_the_compositional_dependence_of_thermal_conductivity_in_nickel_superalloys
The SI unit of electrical resistivity is Ω⋅m. Electrical conductivity (σ) is the reciprocal of electrical resistivity, having the SI unit of siemens per meter (S/m). The σ3/ρ ratio, where ρ stands for density, has S3/kg units. We may consider the following empirical additive linear correlation to predict (estimate) the electrical conductivity of the considered alloy from that of its constituting elements, generically denoted by the index i: σ3/ρ ≈ f1·w1·σ13/ρ1 + ... + fi·wi·σi3/ρi + ... + fn·wn·σn3/ρn. Here, wi express mass ratios for the elements of the alloy, while fi are fitting coefficients that can be adjusted to the available resistivity (and density) data for the alloy composition system. If at least one data point is available, but there is not enough data to determine all these coefficients independently, we may take the fitting coefficients as equal between each other, so that they would reduce to a single fitting coefficient. At least n data points should be available to allow all the coefficients to be determined from the available data.