Data Reduction Based on Local Hausdorff Measures for Forensic Data
Currently, in many domains (e.g. multispectral images, text categorization, biometrics, retrieval of multimedia database, computer forensics), the size of the data sets is so extremely large that real-time systems cannot afford the time and storage requirements to process them. Data reduction techniques are approaches in charge of diminish the quantity of information in order to reduce both memory and execution time. In this paper, the authors proposed a schema to reduce the quantity of instances using local Hausdorff measures. In the schema, they divide the original data set to subsets and use the Hausdorff measures on the subsets, the instances in the set which doesn't change or change the hausdorff distance slightly will be removed, which can reduce the quantity of the original set.