Download PDFOpen PDF in browserCompare of Machine Learning And Deep Learning Approaches for Human Activity RecognitionEasyChair Preprint 14956 pages•Date: September 12, 2019AbstractNowadays analyze of human activity and human behavior can be useful for software design especially for patients. So, human activity recognition is important. The aim of this research was find the best algorithm for human activity recognition. We used Logistic Regression, SVM with RBF kernel; CNN, LSTM, Bi-Directional LSTM and CNN-LSTM algorithms for analyze the data. In the data analyze the accuracy and training time measured and compared. The most accuracy belonged to the CNN-LSTM and Bi-Directional LSTM and the least training time belonged to the SVM with RBF kernel. Keyphrases: Bi-directional LSTM, CNN, CNN-LSTM, Human Activity Recognition, LSTM, SVM with RBF kernel, logistic regression
|