Fusing uncertain data with probabilities

UNCG Author/Contributor (non-UNCG co-authors, if there are any, appear on document)
Alaa Hassan Ahmed (Creator)
The University of North Carolina at Greensboro (UNCG )
Web Site: http://library.uncg.edu/
Fereidoon Sadri

Abstract: Discovering the correct data among the uncertain and possibly conflicting mined data is the main goal of data fusion. The recent research in fusing uncertain data shows that taking source confidence into account helps to achieve this goal because the sources have different degree of accuracy. Thus, understanding different modern fusing techniques and using different data sets can be useful to research community. Previous work has fused uncertain data with and without considering correlation between the sources by using training data sets [5]. In our proposed research, we extended this work by calculating the initial probability which is given by the sources that provide the information and then calculating the final probability for the given data. In our work there is no need to training set in which the algorithm can work with different type of uncertain data sets. Also, we present a method to calculate the threshold of the given data set; and we did our experiments by using two types of data sets; one type contains intentional false and other random false.

Additional Information

Language: English
Date: 2016
Aggressive Approximation, Data Integration, Exact Solution, Fusion Data, Probabilities, Uncertain Data
Data integration (Computer science)
Uncertainty (Information theory)
Probabilities $x Data processing
Data mining

Email this document to