The Study on Packaged Food Shelf life Based on Back-Propagation (BP)
To predict shelf life of food products has been an important issue in the field of food science andpackaging engineering. This article first analyzes the current studies on shelf life prediction, and points outthat these evaluation systems are ineffective for their accuracy and practicability. Then artificial neural networkmethodology is introduced to predict shelf life, using BP network to establish a mathematical predictionmodel. Incorporation of various factors, including food product compositions, package properties, storage andlogistics condition into a single model can be achieved when BP neural network is used. Its self-study andself-adjusting ability make the network can generalize their inherent laws from a large number given dataautomatically. The prediction results are closer to the actual dynamic environment because the BP model ishighly adaptable to the changes of outer circumstances. In this article compared to existing models, BP modelwas proved to be a more simple, adjective and higher accuracy method for predicting shelf life. It will promotethe development trend which to use computer method to cope with the study of food shelf life, and offera new approach and idea for modeling theory of foods shelf life prediction.
Zenghui Sun Lei Zhang
Tianjin University of Science and Technology TUST Tianjin, China
国际会议
The 17th IAPRI World Conference on Packaging(第17届世界包装大会)
天津
英文
410-413
2010-10-12（万方平台首次上网日期，不代表论文的发表时间）