Lax-net: Freezing of Gait Detection in Parkinson's Disease Using LSTM with Attention and XGBoost

Parkinson's disease (PD) is a neurodegenerative disorder marked by motor symptoms like gait freezing (FoG), which notably affects patients' mobility and overall quality of life. Early detection of FOG episodes is crucial for timely intervention and reducing the risk of falls and injuries. In this paper, we present Lax-net, a deep learning-based approach for the early detection of FOG using wearable sensor data with time series information. These data often exhibit complex temporal dependencies, making it challenging to extract meaningful features and accurately classify them. To address this challenge, our proposed method combines the strengths of LSTM with attention mechanisms to capture temporal relationships and identify important features in the data. The derived features are then fed into the XGBoost model, a robust classifier with the ability to grasp non-linear relationships. We evaluate the performance of our proposed method on a benchmark dataset, comparing it with other state-of-the-art methods. The results demonstrate that our combined approach achieves superior accuracy and outperforms individual models such as LSTM and XGBoost.