Random Forest remains one of Data Mining's most enduring ensemble algorithms,\nachieving well-documented levels of accuracy and processing speed, as well as\nregularly appearing in new research. However, with data mining now reaching the\ndomain of hardware-constrained devices such as smartphones and Internet of\nThings (IoT) devices, there is continued need for further research into\nalgorithm efficiency to deliver greater processing speed without sacrificing\naccuracy. Our proposed FastForest algorithm delivers an average 24% increase in\nprocessing speed compared with Random Forest whilst maintaining (and frequently\nexceeding) it on classification accuracy over tests involving 45 datasets.\nFastForest achieves this result through a combination of three optimising\ncomponents - Subsample Aggregating ('Subbagging'), Logarithmic Split-Point\nSampling and Dynamic Restricted Subspacing. Moreover, detailed testing of\nSubbagging sizes has found an optimal scalar delivering a positive mix of\nprocessing performance and accuracy.\n