03-07-2013, 04:26 PM
TUPLE UNIT MINING MODEL
UNIT MINING MODEL.ppt (Size: 1,005 KB / Downloads: 73)
ABSTRACT
To find top k answers in keyword search over relational databases using tuple units.
Improves search efficiency by indexing structural relationship
To answers the keyword query ,multiple tuple units are integrated
two novel indexes
INTRODUCTION
Keyword search over relational databases finds the answers of tuples in the databases which are connected through primary/foreign keys and contain query keywords
Tuple units is a set of highly relevant tuples which contain query keywords
Tuple units can be precomputed and indexed and we can use the indexed tuple unit to efficiently answer keyword query
Use two structure aware indexes SKSA index and KPSA index
Develop new ranking techniques and efficient algorithms to efficiently and progressively find the top k answers
Implement this model in MYSQL. This model achieves high efficiency and result quality
1.DBXplorer: A System for Keyword-Based Search over Relational databases(2002)
DBXplorer returns all rows either from single table or from multiple tables, using FK-joins, such that each row has all the keywords
1.Identify a database and tables and columns within it that are to be enabled for search
2. Create auxiliary tables (Symbol Tables)
1. Look up the Symbol table
2. Searching in possible subsets of tables
3. Construct and execute SQL statement and rank the results before displaying to user
2.BANK(Keyword Searching and Browsing in Databases 2002)
BANKS enables users to extract information in a simple manner without any knowledge of the schema or any need for writing complex queries. BANKS models tuples as nodes in a graph, connected by links induced by foreign key and other relationships. Answers to a query are modeled as rooted trees connecting tuples that match individual keywords in the query.
Proposed System
Proposed system is Tuple unit based method
Which integrates multiple relevant database tuple units to effectively answer keyword queries.
We propose two structure aware indexes and store the structural relationship between different tuple units into indexes
We propose effective ranking techniques to rank the tuple units by taking into account both structural compactness of answers from the database point of view & textual relavancy from the information retrieval point of view
CONCLUSION
We have studied the problem of effective keyword search over relational databases.
We proposed to integrate multiple relevant tuple units to effectively answer keyword queries.
We devised two novel structure-aware indexes, SKSA-Index and KPSA-Index, which incorporate the structural relationships between tuple units and the textual relevancy between input keywords into the indexes.
We proposed a novel ranking mechanism by taking into consideration both the textual relevancy in IR literature and the structural compactness of tuple units from the DB viewpoint
. We have implemented our method, and the experimental results show that our method achieves high performance and outperforms state-of-the-art approaches significantly.