A probabilistic relational model and algebra

Abstract

Although the relational mode] for databases provides a great range of advantages over other data models, It lacks a comprehensive way to handle incomplete and uncertain data. Uncertainty in data values. however, is pervasive in all real-world environments and has i-eceiv[,d muc}] attent)[]n in the literature, Several methods ha}!e been proposed for incorporating uncertain data into relational databases.However, the current approaches have many shnrtcomlngs and havi, not established an acceptable extension af the relational model, In this paper. we propose :1 consistent extension of the relatiana] model. We present a revised relational structure and extend the relational algebra. The extended algebra ]s shown to be closed. N consistent extension of the conventiana] relational algebra, and reducible to tbe latt{,r ('at egorvcs and Subject

Database Support for Probabilistic Attributes and Tuples

“…RELATED WORK Barbará et al [12] and Dey et al [15] proposed the first of the probabilistic models. Building on their work, many robust models for managing tuple uncertainty have been proposed recently.…”

Probabilistic Threshold Range Aggregate Query Processing over Uncertain Data

“…Sarma et al [16] models uncertain data using possible world semantics and a prototype of uncertain data management system, Trio, is developed by the Stanford Info Lab [12]. General issues in modelling and managing uncertain data are addressed by Dey and Sarkar in [4], Lee in [11], and Antova, Koch, and Olteanu in [6]. Querying uncertain data by the probabilistic paradigm has been investigated by Dalvi and Suciu in [2] and Sen and Deshpande in [17].…”

Databases for interval probabilities

“…The fifth operation, conditionalization, is specific to probabilistic databases and represents the procedure of constructing an ESPO containing a conditional probability distribution given an ESPO for some joint probability distribution. First proposed as a database operation by Dey and Sarkar [13] for a relational model with point probabilities, this operation had been extended to non-1NF databases in [10].…”

“…Conditionalization was first considered as an operation of a relational algebra related to a probabilistic data model by Dey and Sarkar [13]; Classical relational algebra has no prototype of it. Intuitively, conditionalization is the operation of computing a conditional probability distribution, given a joint probability distribution.…”

“…Dey and Sarkar [13] introduced a 1NF probabilistic relational model and relational algebra. A tuple in their model is analogous to a single row of a probability table in ours, and their probabilistic relation could contain multiple probability distributions.…”