Positional data for GPS satellites was analyzed within TPOT to identify the optimal classifier algorithm that enables the determination that a GPS satellite is expected to soon maneuver when its orbital X, Y, and Z positional coordinates are used as predictors. The PyOD package was used to confirm that there were maneuvers by the satellites in the datasets (as indicated by the presence of outliers), and that the kNN classifier exhibits a strong grouping relationship between the predictors used in this classification model and the response variable. The results indicate the KNeighbors classifier is an optimal classifier for this type of dataset and inquiry, and that there is a consistent performance difference exhibited between satellite systems. The results indicate that the average performance reported by TPOT’s during the training/testing classification activity was assessed to be significantly higher than the predicted performance when the optimal algorithm was assessed with a similar dataset that was not used in TPOT’s identification of the optimal classifier.