Data Integration has become increasingly important with today's rapid growth of information available on the web and in electronic form. In the past several years, extensive work has been done to make use of the available data from different sources, particularly, in the scientific and medical fields. In our work, we are interested in integrating data from different uncertain sources where data are stored in semistructured databases, markedly XML-based data. This interest in XML-based databases came from the flexibility it provides for storing and exchanging data. Furthermore, we are concerned with reliability of different query answers from various sources and on specifying the source where the data came from (the provenance). In essence, our work lies among three areas of interest, data integration, uncertain databases and lineage or provenance in databases. This thesis extends previous work on information integration to accommodate integration of uncertain data from multiple sources. |