Research And Application Of Technologies For Data Management And Analysis Of Archaeological Excavations

As the archaeometric activities continue to deepen, more and more invaluable data resources generate from archaeological excavations. Data mining technologies which are good at big data analysis can also be applied to massive archaeological data to promote archaeological research. Therefore, it has great value to do research on the integrated management and mining of archaeological excavation data.Firstly, we study the application of classification methods to predict the social class of the departed in a grave. The prediction is done manually nowadays and needs archaeological knowledge which cannot be represented as a set of simple rules. Thus, we apply classification algorithms including decision tree and SVM to learn the implicit knowledge, so that we can predict automatically and correctly. Experiment results show that the use of SVM model and the introduction of domain knowledge to select important funerary objects for classification have achieved good performance.Secondly, we present an association rule mining framework for archaeological data correlation analysis. This framework allows users to preselect their interesting data records and items, so as to do analysis more specifically and avoid the interference of nonrelated data. And the criteria based on interesting measures can effectively filter out some boring frequent item sets and rules. With the actual data collected from LiangZhu cultural site, we mine new implicit correlations among funerary objects, and verify the effectiveness and practicability of the framework.Lastly we describe the design and implementation of the excavated information management system. It provides functions including massive data integration, management, correlation, picture display, and practical data mining tools.