Polygonati Rhizoma varieties and origins traceability based on multivariate data fusion combined with an artificial...

Area of Science:

• Agricultural Science
• Computer Science
• Cheminformatics

Background:

• Polygonati Rhizoma (PR) is a valuable traditional Chinese medicine with varying quality based on origin and variety.
• Accurate identification of PR origin and variety is crucial for quality control and preventing adulteration.

Purpose of the Study:

• To develop and optimize an intelligent algorithm for distinguishing Polygonati Rhizoma varieties and tracing their origins.
• To evaluate the performance of the novel algorithm compared to traditional methods using multidimensional data fusion.

Main Methods:

• Collected multidimensional feature data (spectra, texture, component contents) from three PR species across different origins.
• Applied multivariate statistical analysis to select characteristic factors for origin and variety discrimination.
• Developed and optimized a deep belief network (DBN) classification algorithm combined with multivariate statistical analysis.

Main Results:

• Identified 39 characteristic factors for distinguishing PR origins and 14 for discriminating PR varieties (VIP > 1, P < 0.05).
• The novel AI algorithm achieved 100% discrimination rate for PR varieties.
• The AI algorithm demonstrated 100% accuracy in tracing the origin of PR.

Conclusions:

• The developed AI algorithm based on multidimensional data fusion significantly improves accuracy in variety discrimination and origin tracing of Polygonati Rhizoma.
• This research provides a valuable reference for constructing intelligent algorithms for food and herbal medicine authentication.
• The findings support the application of AI in ensuring the quality and safety of traditional medicines.