Smart wearable sensor-based model for monitoring medication adherence using sheep flock optimization...

Area of Science:

• Biomedical Engineering
• Artificial Intelligence in Healthcare
• Wearable Technology

Background:

• Medication adherence (MA) is critical for effective patient treatment and therapeutic outcomes.
• Sensor technology offers continuous monitoring of patient MA behavior, representing a significant advancement.
• Accurate MA monitoring is essential for improving patient health and reducing healthcare costs.

Purpose of the Study:

• To implement smart wearable sensor devices powered by deep learning (DL) for predicting medication behaviors.
• To develop a novel hand gesture recognition system using wearable sensors for MA prediction.
• To evaluate the effectiveness of advanced DL techniques in analyzing complex MA data patterns.

Main Methods:

• Utilized accelerometer and gyroscope sensors in a smart wearable device to capture hand motion data.
• Developed a deep learning model, the sheep flock optimization algorithm-attention-based bidirectional long short-term memory network (SFOA-Bi-LSTM), for MA behavior classification.
• Employed Z-score normalization for data preprocessing and the SFOA method for hyperparameter optimization of the attention-based Bi-LSTM model.

Main Results:

• The SFOA-Bi-LSTM model demonstrated high performance with 98.90% accuracy, 97.80% recall, 98.80% precision, and 98.62% F1 score.
• Achieved excellent results in predicting medication adherence behavior through hand gesture recognition.
• The model's performance was validated using a five-fold cross-validation approach.

Conclusions:

• The SFOA-Bi-LSTM model shows significant potential for monitoring medication adherence in healthcare applications.
• The model offers advantages in hyperparameter tuning, feature representation, and temporal analysis, outperforming conventional methods.
• The system is robust to noisy data due to effective preprocessing, making it a reliable tool for healthcare professionals.