A deep learning method for children's self-care problems classification using represent learning and focal loss.

Area of Science:

• Pediatric healthcare
• Machine learning applications
• Developmental psychology

Background:

• Accurate diagnosis of children's self-care problems is crucial for healthy development.
• Classifying these problems is complex due to diverse disorders, demanding significant time and effort.
• Existing diagnostic methods may lack the precision needed for effective intervention.

Purpose of the Study:

• To develop an intelligent and precise deep learning model for classifying children's self-care problems.
• To overcome the complexity and time-consuming nature of traditional classification methods.
• To enhance diagnostic accuracy and support timely interventions for affected children.

Main Methods:

• A novel deep learning model comprising two sub-networks was proposed.
• The first sub-network utilized triplet loss for feature representation learning, compressing dimensions and reducing noise.
• The second sub-network employed focal loss to address class imbalance and improve classification accuracy.

Main Results:

• The proposed deep learning model demonstrated superior performance.
• Average accuracy reached 99.78%, with precision, recall, and F1 scores all achieving 0.99.
• The model effectively handled complex classifications and class imbalances.

Conclusions:

• The developed deep learning model achieves state-of-the-art results in classifying children's self-care problems.
• This method offers significant support for the rehabilitation and growth of children with self-care issues.
• The study highlights the potential of deep learning models in advancing pediatric healthcare diagnostics.