Hybrid Sensor Based Fuzzy Clustering Neural Network Classification for Human Activity Recognition
Authors: Ramtin Aminpour and Elmer Dadios
Abstract
IThe smartphone is going to become an all-purpose gadget for the human life and all of them at least armed with accelerometer sensor. In this study, the fuzzy c-means has been considered in the ANFIS model to produce the fuzzy inference system (FIS) to make the classification with the neural network algorithm to detect the six major human activities. The data were taken in real life with the accelerometer sensor of a smartphone. The results of the experiments show that the 97.2% accuracy could be acceptable in the field of study and the clustering structure could make the simulation more robust and faster.