Titlebook: Data Cleaning; Venkatesh Ganti,Anish Das Sarma Book 2013 Springer Nature Switzerland AG 2013

首創(chuàng)精神

Operator: Clustering,at the records in a group be closer to each other, especially to each other than to records in other groups. A custom deduplication task may require that other constraints beyond similarity be satisfied as well. However, closeness to each other by textual similarity is a critical predicate, which ne

公理

Operator: Parsing, records into a target data warehouse often requires the reconciliation of schema of the input records and that of the target records. The process of reconciliation would often involve “segmenting” a column of an input record into multiple target columns. The segmented input records may then be comp

要控制

Task: Record Matching,esent the same real-world entity, often referred to as “matching.” This important task needs to be solved while importing new customer sales records into the customer relation in a data warehouse. The customer records in the incoming sales need to be matched with existing customers to avoid subseque

Nebulous

Proclaim

Conclusion,mponents of the technology. he goals of data cleaning technology in typical enterprise scenarios, as illustrated by the examples in customer and product databases, are to maintain the quality and consistency of data as the data warehouse is either being populated with data for the first time or bein

伙伴

Cougar

Operator: Similarity Join,ied by a textual similarity function which compares the content of the two records. There are a variety of common similarity functions as discussed in the previous chapter. As in record matching, the deduplication task typically involves many predicates. However, a critical one is often based on textual similarity between records.

做作

Data Cleaning Scripts,erators as well as other predicates, which are required for the specific data and domain being considered. Thus, the development of custom data cleaning scripts is expected to be flexible, easy, and efficient all at the same time.

byline

BINGE