K. Lam

Term Weighting in Information Retrieval Using the Term Precision Model

At3STRACT It iS known that the use of weighted, as opposed to binary, content identifiers attached to the records of an information file improves the effectiveness of the retrieval operations Under well-defined conditions the term precision offers the best possible term weighting system A mathematscal model is used in the present study to relate the term precision weights to the frequency of occurrence of the terms in a given document collecuon and to the number of relevant documents a user wishes to retrieve in response to a query This provides for the assignment of user-dependent weights to the content identifiers and relates the term precision weights to other well-known term weighting systems

Optimization of distributed tree queries

Abstract In this paper the problem of finding an optimum strategy of semi joins for solving tree queries is studied under the objective of total time minimization. Tree queries that are conjunctions of equi-join clauses such that any two relations in the query have at most one attribute in common are considered. This class of tree queries is a superset of classes of tree queries, such as chain queries and simple queries, that have been studied for semi-join optimization in the literature. An algorithm based on dynamic programming to find the optimum semi-join strategy for a given query is presented. The search space for finding the optimum is reduced by eliminating strategies that can never be the optimum. This is accomplished by utilizing a set of properties that a potentially optimum strategy should satisfy.

A clustered search algorithm incorporating arbitrary term dependencies

The documents in a database are organized into clusters, where each cluster contains similar documents and a representative of these documents. A user query is compared with all the representatives of the clusters, and on the basis of such comparisons, those clusters having many close neighbors with respect to the query are selected for searching. This paper presents an estimation of the number of close neighbors in a cluster in relation to the given query. The estimation takes into consideration the dependencies between terms. It is demonstrated by experiments that the estimate is accurate and the time to generate the estimate is small.

Adaptive File Allocation in Star Computer Network

In this paper, we study the allocation of files in a star network. Unlike previous algorithms which assume that files are independently accessed and independently assigned, the interaction of files during the processing of queries is directly incorporated into our cost model. We present an adaptive algorithm, which is much faster than existing algorithms on file allocation, obtains solutions which are on the average only 0.1 percent away from the optimal solutions, and possesses many desirable properties such as the satisfaction of some necessary and sufficient conditions for file allocation.

A generalized counter scheme

Abstract The transposition heuristic is a common method used to improve the performance of accessing records in a linked list. After the list has reached a steady state for the transposition heuristic, we begin to keep a frequency count for each record. A method of re-arranging the records, called the generalized counter scheme, is introduced and is shown to be optimal among all possible methods of re-arrangement based on the counts. The scheme is applicable even when the count is small. The usual counter scheme is also shown to be optimal for any finite count.

An approach to probabilistic retrieval

The objective is to relate the effectiveness of retrieval, the fuzzy set concept and the processing of Boolean query. The use of a probabilistic retrieval scheme is motivated. It is found that there is a correspondence between probabilistic retrieval schmes and fuzzy sets. A fuzzy set corresponding to a potentially optimal probabilistic retrieval scheme is obtained. Then the retrieval scheme for the fuzzy set is constructed.