Efficient storage and retrieval of temporal semi structured data / Rasha Abdulaziz Saleh Bin Thalab ; Supervised Osman Hegazy , Mohamed E. Elsharkawi , Neamat Eltazi

Thesis (Ph.D.) - Cairo University - Faculty of Computers and Information - Department of Information system

Interesting research work has recently focused on the problem of retrieving historical data from XML documents, hence temporal XML. Approaches to support temporal XML documents by extending XML or its query languages have been extensively proposed in literature. Unfortunately, these approaches cannot be used efficiently in retrieving this historical data neither by conventional XML indices nor by conventional XML key word search algorithms. The major limitation of these conventional approaches is that they were not built for handling time dimension. In this thesis, we aim at improving historical data retrieval from temporal XML using both structured queries and keyword search. In case of structured queries, we introduce a new index, called TMIX, based on summary structures and coalesced time intervals for temporal data. The new index reduces search space, answers temporal queries efficiently and handles update queries without reconstruction. Extensive performance evaluation has been conducted to evaluate the new index, TMIX, against the state of the art indices. In case of keyword search, we used a graph model to map temporal XML documents objects to real world entities and used life span interval within each object to map the time dimension. We created index structures for keywords and objects for efficient keyword search retrieval. Furthermore, we extended regular ranking models with time to improve retrieval effectiveness. Through extensive evaluation, we show that our proposed temporal approaches outperform traditional XML keyword search methods and improve effectiveness in querying temporal XML documents

Issued also as CD