The reactive mass transfer at spherical gas bubbles in a stationary incompressible fluid is analysed numerically. Only the continuous phase is simulated. The influence of the gaseous phase on mass transfer is neglected. The concentration of the transfer species, which diffuses into the bulk, is assumed to be constant in space and time inside the bubble. Four different reaction types, namely decay, single, parallel consecutive and parallel competitive reaction are evaluated all being irreversible and of first order for the transfer species A. The main parameters for mass transfer are accessed and a parameter study is carried out to quantify their influence. The parameters studied are Re ∈ [2; 100] , Pe ∈ [10E4; 10E6], Da ∈ [0, 10] and different ratios of concentrations and diffusivities, both in the range of [0.1; 10]. The data is then compared with known correlations for non-reactive mass transfer and the enhancement factor. Afterwards regression algorithms from the field of machine learning are implemented and trained on the data sets. The performance of the algorithms is accessed. A trained algorithm is used to predict the mass transfer and the resulting volume change at a rising CO_2 bubble in Water. The effect of conjugated mass transfer is analysed and the results are compared to experimental data.