Efficient and effective aggregate keyword search on relational database

Keyword search on relational databases is useful and popular for many users without technical background. Recently, aggregate keyword search on relational databases was proposed and has attracted interest from both academia and industry. However, two important problems still remain. First, aggregate keyword search can be very costly on large relational databases, partly due to the lack of efficient indexes. Second, finding the top-k answers to an aggregate keyword query has not been addressed systematically, including both the ranking model and the efficient evaluation methods. In this thesis, we tackle the above two problems to improve the efficiency and effec- tiveness of aggregate keyword search on large relational databases. We design indexes efficient both in size and in constructing time. We propose a general ranking model and an efficient ranking algorithm. We also report a systematic performance evaluation using real data sets.