Within the next decade large multi-waveband all sky surveys will become available from telescopes like Euclid, LSST and SKA, and the need for fast analysis of data and images will be of great importance. Here I present the results of running a machine learning code, a Support Vector Machine, on a combined simulated optical and simulated radio catalogue of strong gravitational lenses. When the catalogues are combined we see an increase in the true positive rate for the detection of strong gravitational lenses, and a reduction in the false positive rate. We see that individually the optical catalogues have good detection rates, giving an AUROC score of 0.94 for the simulated KiDS r-band optical and 0.85 for Euclid-VIS band optical. The radio catalogues produce lower scores, with the Gaussian catalogues for SKA phase 1 giving 0.6222 and for SKA phase 2 giving 0.7854. Combining the simulated Euclid VIS-band catalogue with with a simulated SKA radio catalogue containing AGN-like galaxies and we get an AUROC score of 0.9265 for SKA phase 1 and 0.9552 for SKA phase 2. This shows a ~10% improvement on the Euclid-VIS band data alone. We also note that SKA phase 2 will perform better than phase 1, with ~3.5% more detections, and a better true positive to false positive ratio, giving a higher completeness and greater purity across all simulated radio catalogues.