DocumentCode
294148
Title
Knowledge guided data mining
Author
Milne, Dr Robert ; Nelson, Mr Chris
Author_Institution
Intelligent Applications Ltd., Livingston Village, UK
fYear
1995
fDate
34731
Firstpage
42522
Lastpage
42524
Abstract
Intelligent Applications Ltd. have recently completed the final phase of the development of a software model for a client. The client has a very large number of customers made up from members of the general public. The model predicts which customers of the client have an increased likelihood of being lost to a competitor. The model was developed by performing data mining on a representative set of data from the client, using knowledge-guided induction. One of the main benefits to the client is the provision of intelligence, resulting from the data mining, to help them target their mailshot campaigns and other marketing exercises in the most effective way. The model has undergone extensive tests and validation routines to measure its accuracy based on customer data taken from a 3-year time window. The model has consistently yielded extremely high accuracies for most customer types, and more than 90% for certain customer types in particular
Keywords
deductive databases; knowledge acquisition; marketing data processing; accuracy; competition; customer losses; customer types; data mining; knowledge guided data mining; knowledge-guided induction; mailshot campaigns; marketing exercises; software model development; validation routines;
fLanguage
English
Publisher
iet
Conference_Titel
Knowledge Discovery in Databases, IEE Colloquium on (Digest No. 1995/021 (A))
Conference_Location
London
Type
conf
DOI
10.1049/ic:19950117
Filename
476229
Link To Document