A text feature extraction model for hazardous chemical recovery identification and attribute classification embedded...

Area of Science:

• Environmental Science
• Chemical Engineering
• Data Science

Background:

• Industrial production and research frequently use hazardous chemicals, necessitating effective recovery and hazard characterization for environmental health.
• Current table-query methods for analyzing hazardous chemicals are inefficient, highlighting the need for advanced technologies like big data.

Purpose of the Study:

• To improve the identification and understanding of hazardous chemical characteristics and risks.
• To develop an automated model for recognizing hazardous chemical information, reducing subjective errors.

Main Methods:

• Incorporation of a hazardous chemicals domain knowledge graph into risk identification.
• Construction of a hazardous chemical recovery text recognition model named KG-TextRCNN.

Main Results:

• The KG-TextRCNN model achieved an average precision of 99.30% for text corpus recognition, with recall and F1 scores above 99%.
• The classification step using max pooling yielded average precision, recall, and F1 scores of 85.19%, 86.29%, and 85.27%, respectively.
• The model demonstrated stable overall performance with high precision and recall rates across most categories.

Conclusions:

• The developed KG-TextRCNN model effectively identifies hazardous chemicals, offering a significant improvement over traditional methods.
• The integration of knowledge graphs and deep learning enhances the accuracy and reliability of hazardous chemical risk assessment.