Data Mining Analytics for BI and Decision Support

Read this article to learn how data mining employs techniques from statistics, pattern recognition, and machine learning to support decision making. The article further provides clarification to the other key algorithm families, such as predictive modeling, and segmentation used in data mining.

Background

We are seeing today widespread and explosive use of database technology to manage large volumes of business data. The use of database systems in supporting applications that employ query based report generation continues to be the main traditional use of this technology. However, the size and volume of data being managed raises new and interesting issues. Can we utilize methods wherein the data can help businesses achieve competitive advantage, can the data be used to model underlying business processes, and can we gain insights from the data to help improve business processes? These are the goals of Business Intelligence (BI) systems, and Data Mining is the set of embeddable (in BI systems) analytic methods that provide the capabilities to explore, summarize, and model the data. Before applying these methods to data, the data has to be typically organized into history repositories, known as data warehouses. Data warehousing may require integration of multiple sources of data, which may involve dealing with multiple formats, multiple database systems, and distributed databases, cleaning the data, and creating unified logical view of the underlying non-homogeneous data.

Online Analytics Processing (OLAP) is an extension of Structured Query Language (SQL) framework to accommodate queries that would otherwise have been computationally impossible on a relational database management system. This is achieved by utilizing and storing pre-computed aggregates(e.g.credit-card sales in a certain geographic region over a certain time period) that are automatically updated as the underlying data changes. Deciding upon which aggregates to pre-compute is determined by the business end-user. Providing technical capabilities for automatic computation and updating of aggregates is strength if OLAP analytics. Data mining analytics try to go beyond OLAP by providing abilities for discovering insights that are computer driven and not end-user driven. Data size is increasing at a rate far exceeding any rates that end-users can cope with. Providing solutions when end-users cannot reasonably supply all possible aggregates to pre-compute, or when it is not possible to express an insight as a pre-computed aggregate, is the goal of data mining analytics.


Source: A.Venkateswara Rao, G. L. Aruna Kumari, and Mandava V. Basaveswara Rao, http://www.computerscijournal.org/vol1no2/data-mining-analytics-for-business-intelligence-and-decision-support/
Creative Commons License This work is licensed under a Creative Commons Attribution 4.0 License.