BSDGAN: Balancing Sensor Data Generative Adversarial Networks for Human Activity Recognition

The development of IoT technology has enabled various sensors to be integrated into mobile devices. Human Activity Recognition (HAR) based on sensor data has become an important part of the field of machine learning and ubiquitous computing. However, human activities are not evenly distributed because of the rarely performed activities. The dataset is extremely imbalanced. In this paper, we propose Balancing Sensor Data Generative Adversarial Networks (BSDGAN) to generate sensor data for rarely performed human activities. To address extreme imbalances in human activity dataset, an autoencoder is employed to initialize the training process of BSDGAN, which to ensure the data features of each activity can be learned. Add the generated activity data to the original data to generate a new dataset that increases the proportion of rarely performed human activities, which makes the dataset balanced. We deployed multiple human activity recognition models on two publicly available imbalanced human activity datasets, WISDM and UNIMIB. Experimental results show that the proposed BSDGAN can effectively capture the data features of real human activity sensor data, and generate realistic synthetic sensor data. Meanwhile, the balanced activity dataset is effective for the activity recognition model to improve the recognition accuracy.