An Overview of Data Mining

Data mining tools allow the user to sift through the large amount of data stored in data warehouses.

Capability of data mining includes:

·         Automated prediction of trends and behaviours eg. Targeting a market for promotional mailing lists

·         Automated discovery of previously unknown pattern eg. Pattern discovery detects fraudulent credit card transactions and identify anomalous data entries

·         Provides architecture, and tools for business executives to systematically organise, understand and use their data to make strategic decisions.

·         Allows integration of a variety of applications such as spreadsheet and other development tools such as Gentia , so that data can be analysed and processed.

·         It can integrate multiple heterogenous sources such as from data marts, data warehouses, in other database stores on the internet and intranet servers

The architecture of a typical data mining system consists of:

·         Data warehouse - where data cleaning and integration techniques are performed on data.

·         Data warehouse server - fetches the relevant data based on data mining request

·         Knowledge base - the domain knowledge that guides the search, evaluates the interestingness of the patterns and includes concept hierarchies and metadata.

·         Data mining engine - helps the user find the right kind of patterns (either descriptive or predictive mining tasks) based on these patterns, classes are derived and functions are performed. Click here for a description of data mining functionalities

·         Pattern evaluation module - employs interestingness measures and focus the search towards patterns

·         GUI - the Graphical User Interface allows the communication between the user and the data mining system. It receives the data-mining query; Click here for a technical description of DMQL Data Mining Query Language. The GUI allows the user to browse the data warehouse schema or data structures, evaluate the mined pattern and visualise patterns in different forms.

Data mining has many practical implications in E-commerce, as described in the functionalities of data mining, the numerous techniques and tools available enables the user to extract important information for marketing by sales analysis, security by fraud detection, discovering new target markets with cluster analysis, describing consumer behaviour and trends with evolution analysis, and estimating future values with forecasting tools, association and sequencing techniques which use pattern matching to identify relationships between events  at one time or over a period respectively, and more recently the gathering of user inputs in web pages and form fields with web mining technology.

Click here for more technical information on DMQL

Click here for more technical information on Data Mining Functionalities

For more information on data mining, go to the critiques section, where a selection of the best data mining websites are linked and reviewed.

Information on this page are based on Data Mining & Datawarehousing, Farhad Daneshgar, 2001.

Any enquiries email [email protected]