Provided by: International Journal of Emerging Technology in Computer Science and Electronics ( IJETCSE)
Topic: Data Management
Date Added: Jul 2014
Conventional feature selection classifiers work with known and precise data values. In recent data collection methods, appreciable amount of attributes are uncertain. The uncertain attributes, in almost all applications, have more influences on the data set on information classification and feature selection constructs. Uncertainty needs to be handled properly. Reasons for uncertainty are due to measurement errors, quantization errors, data staleness and multiple repeated measurements. Uncertainty of a data item is represented in terms of multiple values. Usually uncertain data are abstracted by statistical derivatives (e.g., mean, standard deviation, median etc.).