Human activity recognition using CNN-BiLSTM-LightGBM hybrid model

With the rise of wearable devices and IoT technologies, classifying and understanding human activities has become an important research area. In this study, a hybrid approach that combines deep learning and machine learning methods is proposed to enhance the potential of wearable technologies in tracking human activities and provide more effective solutions to activity recognition problems. The proposed hybrid approach, consisting of Convolutional Neural Networks (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and Light Gradient Boosting Machines (LightGBM), performs feature extraction through a combined CNN-BiLSTM model and utilizes the Light Gradient Boosting Machine algorithm for the classification process. A dataset containing human activities captured from accelerometer and gyroscope sensors is used for experiments. The experimental results demonstrate that the proposed CNN-BiLSTM-LightGBM approach achieves a remarkable accuracy rate of 99.51% and an impressive F1 score of 99.50% in classifying human activities.