HADCLEAN: A hybrid approach to data cleaning in data warehouses

Hybrid Approaches for Data Cleaning in Data Warehouse

“…The modified version of PNRS [9] (Near Miss Strategy) will accept the raw data as the input and clean the data based on some dictionary database. This process will get the data by applying the steps, those are adding, deleting or updating the data, removing some blank spaces and sometime inter changing the two letters.…”

“…This approach [9] is having lot of techniques for grouping the words. This is used as the priority of the approaches for cleaning the data.…”

“…In Transitive Closure, [9] matching of records is done on only one key (attribute) match. It also needs manual intervention to decide whether there is any duplication or correction in the records.…”

“…But in Modified version of Transitive Closure more than one key is used for matching the records into one group. The approach [9] is described as follows: prioritizing the keys in the Transitive Closure Algorithm at 2 levels. At First level, keys are divided into 3 categories: 1) Primary: unique for a person (either one-to-one or one-to-many) 2) Secondary: relatively unique 3) Tertiary: not so unique At second level, inside the categories, order of the keys is based on decreasing priority of uniqueness/importance.…”

“…This approach [9] is mainly working on the two algorithms; they are a) Near Miss Strategy and b) Phonetic Algorithm.…”