A data mining definition can vary based on history and changes. You’ll often see it mixed in with business intelligence, business analytics, and predictive analytics. Some other definitions focus on the statistical end of the process. It’s an interesting concept in that it truly does combine elements of all of these fields. For a successful data mining implementation, all of these areas would be accessed.
Since the concept has been cobbled together from separate sources, there isn’t a historical or scientific definition for the idea. Even the Data Mining Group (DMG) does not have a clear cut definition. Data mining is often defined as part of Knowledge Discovery process.
It’s important to understand the terminology as it can be a popular concept to consider in the marketplace, especially with big data and social networking and data like Facebook. Companies have gotten good at accumulating data, but this isn’t the true definition. Data has to be gathered, cleansed, modeled, and analyzed before you can really call it being “mined”. Often times this needs to be repeated, so it’s not a one time run.
You’ll often find many where simply analyzing a large set of data is labeled as data mining. A large set of reports run against a big datawarehouse may be a good example of enterprise reporting, but not mining. In general, business intelligence is a way to view past history, using typical reports, dashboards, or documents. Data mining works to unveal future data patterns and predictions. However, just having a large dataset doesn’t make data mining. It also needs to include predictive and statistical processes around the data.
The most encompassing definition that I know defines data mining as the following: Data mining is a process to use automation to analyze data sets, using datawarehouse technology, business intelligence, predictive modeling, and other methodology, to predict or uncover patterns and future activity in the data.
Copyright 2011-2017 BusinessIntelligenceBase.com