Identification Methods Based On Modern Nuclear Analytical Techniques Of The Ancient Jun Porcelain

The Jun Kiln site located at Yuzhou city in Henan Province. The coppe red galze was made successfully at Jun Kiln. Heavily loaded with historical and cultural information, Ancient Jun Porcelain has played a very important and unique role throughout China's long history of porcelain production. The identification study of Jun official porcelain and Jun folk porcelain is quite a knotty problem due to the diversity of samples.In this dissertation we selected the porcelain shreds excavated in Juntai official Kiln of Yuzhou city, Qingliangsi official Kiln of Baofeng city and Ru folk Kiln as the objects of interests. Their chemical composition and content were studied by Proton Induced X-Ray Emission (PEXE), Instrumental Neutron Activation Analysis (NAA) and X-Ray fluorescence (XRF); we analyzed the tested data by using factor analysis and cluster analysis of multivariate statistical analysis; and we also did pattern recognition of the data by using support vector machines of Statistical Learning Theory, the results showed that:1. It is not the same of the body and glaze's sources and the formula of Jun official porcelain and Ru official porcelain; Jun official porcelain and Ru folk porcelain's glaze sources and formula is completely different;2. It can be well identification Jun official glaze samples and Ru glaze samples;3. Parts of Ru folk porcelain samples and Jun official porcelain samples are difficult to distinction;4. SVM method is more suitable for handling small samples, such as ceramics data, it be able to overcome overfitting and underfitting of the data processing, the result is even better than factor analysis and cluster analysis's results;5. The multivariate statistical analysis and the support vector machines each have their own advantages, if we took each other's advantages when analyzing, the result can be more detailed and accurately;6. Modern nuclear analytical techniques, multivariate statistical analysis and support vector machine technology, provide a proven, accurate and reliable method to correctly identify the ancient porcelains.