Abstract:
A pair wise comparison of data objects is a requisite step in many data mining applications, but has quadratic complexity. In applications such as record linkage, blockin...Show MoreMetadata
Abstract:
A pair wise comparison of data objects is a requisite step in many data mining applications, but has quadratic complexity. In applications such as record linkage, blocking methods may be applied to reduce the cost. That is, the data is first partitioned into a set of blocks, and pair wise comparisons computed for pairs within each block. To date, blocking methods have required the blocking scheme be given, or the provision of training data enabling supervised learning algorithms to determine a blocking scheme. In either case, a domain expert is required. This paper develops an unsupervised method for learning a blocking scheme for tabular data sets. The method is divided into two phases. First, a weakly labeled training set is generated automatically in time linear in the number of records of the entire dataset. The second phase casts blocking key discovery as a Fisher feature selection problem. The approach is compared to a state-of-the-art supervised blocking key discovery algorithm on three real-world databases and achieves favorable results.
Published in: 2013 IEEE 13th International Conference on Data Mining
Date of Conference: 07-10 December 2013
Date Added to IEEE Xplore: 03 February 2014
Electronic ISBN:978-0-7695-5108-1