A new dataset for human activity recognition and its classification with deep learning models

recent years, the use of mobile sensors for human activity recognition has become an intriguing researcharea due to the proliferation of wearable and mobile sensors. In Muslim life, the prayer (Salah) is an activitythat believers are obligated to perform five times a day. In this study, a new dataset, including Salah, ispresented for use in human activity recognition. Named HAR-P (Human Activity Recognition for Praying),the dataset comprises linear acceleration, acceleration, magnetic field, and gyroscope sensor data for eightactivities: walking, running, typing, downstairs, upstairs, sitting, standing, and praying. Data were collectedfrom 50 male volunteers aged 15-60 using a smartwatch for the HAR-P dataset. The classificationperformance of LSTM, ConvLSTM, and CNN-LSTM models was compared for the HAR-P dataset. The highest average classification accuracy of 91% was achieved with the LSTM method using linearacceleration sensor data and the ConvLSTM model using acceleration sensor data, while the lowest averageaccuracy of 83.6% was attained with the gyroscope sensor data and the ConvLSTM method.