Rise in sensor-based devices has made sensor data cheap and easily available from various sensors. In most cases, sensor data are in the form of converted amplitude/voltage data that essentially become continuous data. Iterative Dichotomiser 3 (ID3) is the basis algorithm used for decision trees. Decision tree algorithms are known to be best suited for discrete data. The accuracy of ID3 algorithm can be improved by converting continuous data to discrete form using some sort of binning/quantization technique. This research study uses clustering technique for quantization of continuous data obtained from accelerometer and gyroscope sensor. The continuous sensor data are divided into clusters, generating discrete data in the form of centroids of each cluster. The obtained discrete data are used for performance benchmarking of ID3 algorithm against statistical binning technique of discretization as well as original continuous data. Our research study shows that the accuracy of ID3 with clustering-based discretization overperforms other techniques by 2–3% improvement in accuracy.