S. Shekhar, B. Hamidzadeh and A. Kohli. `Learning transformation rules for semantic query optimisation: a data-driven approach', IEEE, 949-964, 1993.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....convincing arguments exist for their use[31] Fundamentally different from this use of declared constraints is that of discovering infor4 mation about the instance itself that can be used in SQO. For example, several researchers have incorporated rules discovered from the data to rewrite queries [1, 10, 32, 39]. For example, Bell [1] uses discovered FDs to eliminate group by and distinct operations. Recently, work by Godfrey et al. 6, 7] demonstrates that instance knowledge yields significant, positive results in SQO. They use the concept of a soft constraint (SC) which reflects knowledge about the ....

Shekhar S., Hamidzadeh B., Kohli A., and Coyle M. Learning transformation rules for semantic query optimization: a data-driven approach. IEEE Transactions on Knowledge and Data Engineering 5, 6 (December 1993), 950--964.

....in a very expressive way. The knowledge represented as a FDT is understandable and it differs from black box systems as neural networks. Moreover, a FDT is equivalent to a set of fuzzy rules [6] And such kind of induced rules can be introduced to optimize the query process of the database [7, 33] or to deduce decisions from data [1, 2, 15, 16] FDTs enable us to obtain various kinds of such rules [19] Thus, it is a powerful knowledge representation. A FDT, as a set of fuzzy rules, can be used as knowledge base to help flexibly querying a database. Nowadays, literature related to FDT ....

....rule base induced as a FDT is a new form of knowledge that can be associated with the database. This knowledge can be used in different ways. First of all, it enables us to improve the querying process of the database. A set of rules can be introduced to optimize the query process of the database [7, 33] or to deduce decisions from data [1, 2, 15, 16] Such a set of induced fuzzy rules can be associated with a database as a knowledge base that can be used to help answering frequent queries. A fast response can be found for a query on the value of an attribute. It can also lower the conditions on ....

S. Shekhar, B. Hamidzadeh, A. Kohli, and M. Coyle. Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engineering, 5(6):950 964, December 1993.

....to induce knowledge from such databases. A particular instance of such knowledge is represented by fuzzy knowledge. We study induction of a set of fuzzy rules from a database, the inductive learning scheme. Such kind of induced rules can be introduced to optimize the query process of the database [8, 16] or to deduce decisions from data [2, 3, 11, 12] Many works have been done on the topics of inducing rules from database, but the introduction of fuzzy set theory in the process of induction of rules is more recent. In the induction of knowledge from data, difficulties appear when considering ....

S. Shekhar, B. Hamidzadeh, A. Kohli, and M. Coyle. Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engineering, 5(6):950 964, December 1993.

....to each query. But this can provide a random sample with low probability of containing useful rules. The histogram based rule sets described in this paper are small, easily accessed, and all rules have a good chance of being used. Previous semantic optimisation algorithms [e.g. Hs 95, Ki 81, Sh 88, Si 88] have been iterative, progressively applying new rules as the query is changed by previous rules. Rule application means adding consequent conditions to the query tom rules whose antecedents are implied by query conditions) This is slow. Its sequential character is tindesirable in a ....

....[13. 30] D Branches also occur when the same antecedent condition appears in more than one rule, whose consequents describe different attributes. Eg: a = PC 49 ) 115.04 b 208.63) and (a = PC 49 ) 3.6.80 c 12.7. 82) The traditional approach to semantic query reformulation [e.g. Sh 88] does not refer to a graph. It entails iteratively adding consequent conditions tom rules to the query if the rule antecedents are implied either by Original Conditions in the query or by conditions subsequently added to the query. The process, known as Semantic Expansion, can be seen as path ....

S. Shekhar, B. Hamidzadeh, A. Kohli, and M. Coyle. Learning transformation rules for semantic query optimization: A data-driven approach, IEEE Transactions on Knowledge and Data Engineering, 5(6), 1993, pp 950-964.

....Extracting semantic information from database schemas and contents, often called rule discovery, has been studied over the last several years. Rules can be inferred from integrity constraints [3, 2, 24] or can be discovered from database content using machine learning or data mining approaches [5, 7, 10, 21, 22, 24]. It has also been suggested that such rules be used for query optimization [11, 21, 22, 24] in a similar way that traditional integrity constraints are used in semantic query optimization [4, 15, 6] Many algorithms for mining functional dependencies, which can be considered a special type of ....

....has been studied over the last several years. Rules can be inferred from integrity constraints [3, 2, 24] or can be discovered from database content using machine learning or data mining approaches [5, 7, 10, 21, 22, 24] It has also been suggested that such rules be used for query optimization [11, 21, 22, 24] in a similar way that traditional integrity constraints are used in semantic query optimization [4, 15, 6] Many algorithms for mining functional dependencies, which can be considered a special type of check constraints, have been developed over the last years [12, 1, 17, 20] A lot of work has ....

S. Shekar, B. Hamidzadeh, A. Kohli, and M. Coyle. Learning transformation rules for semantic query optimization. TKDE, 5(6):950--964, Dec. 1993.

....Extracting semantic information from database schemas and contents, often called rule discovery, has been studied over the last several years. Rules can be inferred from integrity constraints [2, 3, 30] or can be discovered from database content using machine learning or data mining approaches [5, 7, 12, 27, 28, 30]. It has also been suggested that such rules be used for query optimization [13, 27, 28, 30] in a similar way that traditional integrity constraints are used in semantic query optimization [4, 6, 17] A lot of work has been devoted to the problem of estimating the size of the result of a query ....

....has been studied over the last several years. Rules can be inferred from integrity constraints [2, 3, 30] or can be discovered from database content using machine learning or data mining approaches [5, 7, 12, 27, 28, 30] It has also been suggested that such rules be used for query optimization [13, 27, 28, 30] in a similar way that traditional integrity constraints are used in semantic query optimization [4, 6, 17] A lot of work has been devoted to the problem of estimating the size of the result of a query expression. Approaches based on sampling were explored in [11, 18] and on histograms in [15, ....

S. Shekar, B. Hamidzadeh, A. Kohli, and M. Coyle. Learning transformation rules for semantic query optimization. TKDE, 5(6):950--964, Dec. 1993.

....all rules from a given query and databases. The approaches may be classified as heuristic based systems [Siegel et al. 1992] logic based systems [Chakravarthy et al. 1990] graph based systems [Shenoy and Ozsoyoglu, 1989] and data driven systems [Hsu and Knoblock, 1994; Lowden et al. 1995; Shekhar et al. 1993] However, having a large rules set remains a problem in all the existing systems since rules are produced automatically regardless of how effective they might be in the query transformation process [Chan and Wong, 1991; Han et al. 1993; Piatetsky Shapiro and Matheus, 1993; Savnik and Flach, ....

S. Shekhar, B. Hamidzadeh and A. Kohli. Learning transformation rules for semantic query optimisation: a data-driven approach. IEEE, 949-964, 1993.

....produces summary information that can either answer a query without consulting the data itself, or else modify the query to a form the data server will be able to process more quickly. This query modification operation using knowledge of the data is known as Semantic Query Optimisation (SQO) [1, 3, 5,7]. Relational data is discussed in this paper. This application requires continuous data analysis, as the focus of query interest in the database changes with time. The analyser s activity is guided partly by information it discovers during examination of the data, and partly from query ....

....useful, but examines the data directly to test whether the rules are supported by the data. We now generalise the queryrecommended rule to a pattern for a set of rules to be obtained by data analysis, which then examines and shortlists the most probably useful subset for future queries. Shekhar [ 5 ] used a classification grid to examine the distribution of data values and derive rules with a wide range of structures, whose applicability to a wide range of queries, robustness against data changes, ease of maintenance, and access speed is poor compared with attribute pair rules. Our data ....

S. Shekhar, B. Hamidzadeh, A. Kohli, M. Coyle, Learning Transformation Rules for Semantic Query Optimization: A Data-Driven Approach, IEEE Trans Data and Knowledge Engineering 5(6) 950-964, 1993.

....towards scaling up to terabyte data sets. We propose to develop scalable algorithms for improving the spatial accuracy of PLUMS derived models, as well as implement and evaluate them in this proposal. In addition, we have been working on developing effective and scalable data mining techniques [53, 56] spatial data models [55] spatial indexing [61] and spatial query processing strategies [55, 56] for different application domains, including ecology [6] transportation [55, 59] and terrain visualization [62] Our team is capable of addressing the proposed research issues. 4 Proposed Work Our ....

....the spatial accuracy of PLUMS derived models, as well as implement and evaluate them in this proposal. In addition, we have been working on developing effective and scalable data mining techniques [53, 56] spatial data models [55] spatial indexing [61] and spatial query processing strategies [55, 56] for different application domains, including ecology [6] transportation [55, 59] and terrain visualization [62] Our team is capable of addressing the proposed research issues. 4 Proposed Work Our preliminary work has shown that simple implementation of PLUMS achieves comparable spatial ....

S. Shekhar and B. Hamidzadeh. Learning transformations rules for semantic query optimization: A data-driven approach. IEEE Trans. on Knowledge and Data Eng. (Special Issue on Discovery in Databases), 5(6), 1993.

....received is lower than the charge: select INSURANCECARRIER, PAYMENTRECEIVED, CHARGE from CLAIMSTABLE where PAYMENTRECEIVED CHARGE The DBMS interface is where database queries are generated. This operation can be done without intelligence, although recent research (see [Han et al. 1993] and [Shekhar et al. 1993] in this issue) has shown how discovery techniques can improve the performance and results of database queries. In our model, the DBMS interface plays a subordinate but important role. It can be argued that a system that does discovery on databases must be able to access a database. From a ....

Shashi Shekhar, Babak Hamidzadeh, Ashim Kohli, and Mark Coyle. Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engineering, to appear, 1993.

....database state may become invalid or inconsistent with a new database state. Many applications of data mining and knowledge discovery require discovered knowledge to be consistent in all database states. Examples include rule discovery for semantic query optimization (SQO) Siegel et al. 1991, Shekhar et al. 1993, Hsu and Knoblock, 1994, Hsu and Knoblock, 1996b, Hsu, 1996) learning an integrated ontology of heterogeneous databases (Dao and Perry, 1995, Ambite and Knoblock, 1995) functional dependency discovery (Mannila and Raiha, 1994, Bell, 1995) knowledge discovery for decision support, etc. ....

Shekhar, S., Hamidzadeh, B., Kohli, A., and Coyle, M. 1993. Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engineering 5(6):950--964.

....robust against database changes. This way, the learner will be able to learn effective and robust rules for semantic query optimization. 1. 4 Closely Related Work Previously, three approaches to automating the knowledge acquisition for semantic query optimization were proposed in [Siegel, 1988] [Shekhar et al. 1993] and [Yu and Sun, 1989] The first approach is a query driven approach due to [Siegel, 1988, Siegel, 1989] Siegel s system learns simple rules, a limited form of rules that allows exactly one literal on each side of implication. His system uses a set of predefined heuristics combined with example ....

....limitation of this approach is that the heuristics depend only on the query specification but do not take properties of data into account, and thus may miss learning many high utility rules. Shekhar et al. develop a data driven approach to the learning problem of semantic query optimization [Shekhar et al. 1993]. Their system is based on an assumption that useful semantic rules can be derived from the nonuniform distributions of attribute values. To detect nonuniform data distributions, their system constructs a set of data distribution grids, such as the one appears in Table 1.3, adopted from [Shekhar ....

[Article contains additional citation context not shown here]

Shashi Shekhar, Babak Hamidzadeh, Ashim Kohli, and Mark Coyle. Learning transformation rules for semantic query optimization: A datadriven approach. IEEE Transactions on Knowledge and Data Engineering, 5(6):950--964, 1993.

....to the database since changes cannot violate the integrity constraints. Unfortunately, scheme based constraints are typically so general that they are of little use. For example, an integrity constraint may require an employee s salary to be greater than zero. Recently, a number of researchers [YS89, SSS92, SHKC93, HK94] have proposed methods for discovering instance based constraints (also referred to as dynamic constraints in [YS89] and derived constraints in [SSS92] which are only valid for particular instances of the database. Instance based constraints contain more information than scheme based constraints ....

....be some constraints relating the two queries. This is a type of query driven approach that requires two similar queries before any constraints are discovered. In addition, this method requires that the results of previous queries be stored and matched against future queries. Data Driven Methods [SHKC93] has proposed a data driven approach that uses grid files to inspect combinations of attribute values for a given data set. The zeros in the grid file indicate constraints. The advantage of this approach is that constraints can be found regardless of the query history. However, since it is ....

S. Shekhar, B. Hamidzadeh, A. Kohli, and M. Coyle. Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engineering, 5(6):950--964, 1993.

....saved 30.99 39.14 1.31 125.46 Gain of total elapsed time 57.1 87.8 12.9 59.1 Average overhead 0.08 0.07 0.07 0.11 Times rule fired 5.00 6.00 4.18 7. 00 6 RELATED WORK Previously, two systems that learn background knowledge for semantic query optimization were proposed by (Siegel 1988) and by (Shekhar et al. 1993). Siegel s system uses predefined heuristics to drive learning by an example query. This approach is limited because the heuristics are unlikely to be comprehensive enough to detect missing rules for various queries and databases. Shekhar s system is a data driven approach which assumes that a set ....

Shekhar, S.; Hamidzadeh, B.; Kohli, A.; and Coyle, M. (1993). Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engineering 5(6):950--964.

....of multidatabase systems because it minimizes the dependency on how individual sources execute a query. When a new information source is integrated into the system, the optimizer can still be used with minimal modification. Many algorithms are available for learning useful semantic knowledge [16, 22, 19, 23, 24, 25]. 1.1 Query Plans A query plan is a directed acyclic graph with its nodes as plan steps and its edges as the ordering constraints that specify data flow direction as well as the order in which the plan steps should be executed. Query plans generated by existing multidatabase query processing ....

....the original query because the system saves the time for the unnecessary comparisons. Semantic query optimization is not widely used in practice largely because it is difficult to encode useful semantic knowledge. Several algorithms have been designed specifically for this purpose previously [16, 22, 19]. Semantic knowledge used in our experiment is learned automatically by our knowledge discovery system [23, 24, 25] The optimization approach described in this paper extends previous work in SQO, which focuses mainly on conjunctive queries in a stand alone database. This approach provides global ....

S. Shekhar, B. Hamidzadeh, A. Kohli, and M. Coyle, "Learning transformation rules for semantic query optimization: A data-driven approach," IEEE Transactions on Knowledge and Data Engineering, vol. 5, no. 6, pp. 950--964, 1993.

....cost. 2 A set of high utility semantic rules is crucial to the performance of a semantic query optimizer. Since it is difficult to encode sufficient semantic rules, researchers have proposed several approaches to rule induction for semantic query optimization (Siegel 1988; Yu Sun 1989; Shekhar et al. 1993; Hsu Knoblock 1994) A rule maintenance approach is also necessary because the learned rules may become inconsistent with data after updates to the database, and the number of rules may grow so large that they may slow down the optimization and reduce the savings. Therefore, a complete semantic ....

Shekhar, S.; Hamidzadeh, B.; Kohli, A.; and Coyle, M. 1993. Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engineering 5(6):950--964.

...., R213, R212, R212 200000 ) member( R213, APFD , ADLS , WMY2 , NPTU ] Table 1: Example rules learned from a database states. Examples include rule discovery for semantic query optimization [ Hsu, 1996, Hsu and Knoblock, 1994, Hsu and Knoblock, 1996b, Siegel, 1988, Siegel et al. 1991, Shekhar et al. 1993 ] learning an integrated ontology of heterogeneous databases [ Dao and Perry, 1995, Ambite and Knoblock, 1995 ] functional dependency discovery [ Bell, 1995, Mannila and Raiha, 1994 ] knowledge discovery for decision support, etc. However, most approaches to these problems assume static ....

Shashi Shekhar, Babak Hamidzadeh, Ashim Kohli, and Mark Coyle. Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engineering, 5(6):950--964, 1993.

.... saving of this class is 43.48 percent, comparable to the SQO systems using hand coded rules (King 1981; Shekhar et al. 1988; Shenoy and Ozsoyoglu 1989) 17.6 Related Work Previously, systems for learning background knowledge for semantic query optimization were proposed by Siegel (1988) and by Shekhar et al. 1993). Siegel s system uses predefined heuristics and an example query to drive the learning. This approach is limited because the heuristics are unlikely to be comprehensive enough to detect missing rules for various queries and databases. Shekhar et al. s system uses a data driven approach which ....

Shekhar, S., Hamidzadeh, B., Kohli, A., and Coyle, M. 1993. Learning Transformation Rules for Semantic Query Optimization: A Data-Driven Approach. IEEE Transactions on Knowledge and Data Engineering 5(6):950--964.

....The existence of many potential spatial relationships implies that consistency checking in the context of spatial databases is more challenging vis a vis its traditional relational counterpart. Consistency checking can also be used for spatial query processing via semantic query optimization[21] and reasoning for spatial qualitative relationships. For example, given a spatial query S and a set of of spatial constraints SC, if the query S is inconsistent with respect to SC then the answer to S is null. Currently most of consistency checking is based on Allens s algorithm [3] which was ....

S. Shekhar and B. Hamidzadeh. Learning Transformation Rules for Semantic Query Optimization: A Data-Driven Approach. IEEE Trans. Knowledge and Data Eng.(Spatial Issue on Discovery in Databases), October 1993.

....framework can help in developing new insights and algorithms for discovering other patterns such as association rules, clusterings and sequences. Another area of interest is in exploring specific data mining application domains such as database integration [GSR96] semantic query optimization [SHKC93] and generalization [AS92] 7 Acknowledgments We would like to thank Dr. Arun Swami from Silicon Graphics Inc. and Dr. Bamshad Mobasher for their valuable contributions in the various stages of the project. We would also like to thank the data mining group at the Department of Computer Science ....

S. Shekhar, B. Hamidzadeh, A. Kohli, and M. Coyle. Learning transformation rules for semantic query optimization: A data-driven approach. IEEE Transactions on Knowledge and Data Engg., 5(6):950--964, December 1993.

No context found.

S. Shekhar, B. Hamidzadeh and A. Kohli. `Learning transformation rules for semantic query optimisation: a data-driven approach', IEEE, 949-964, 1993.

No context found.

S. Shekhar, B. Hamidzadeh and A. Kohli. "Learning transformation rules for semantic query optimisation: a data-driven approach", IEEE, 1993, pp. 949-964.

No context found.

S. Shekhar and B. Hamidzadeh. Learning Transformations Rules for Semantic Query Optimization: A Data-Driven Approach. IEEE Trans. On Knowledge and Data Eng. (Special Issue on Discovery in Databases), 5(6), 1993.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC