Business Intelligence typically answers questions in what-if scenarios. Data mining shows patterns and hidden relationships just by looking at the data. The more granular and detailed the data in the domain of analysis, the better is the capability of data mining tools to perform. Predictive modeling is heavily dependent on formulae comprising of relationships among various predictors carefully chosen by business rules and based on validation against available facts.
Visualization is the process of displaying large volume of data points in a human comprehensive form. It brings out clusters of similar information together and also shows the values in various visual formats like colors and shades that make it easy for non-technical business decision makers to comprehend the data.
While data mining and predictive analytics is a powerful method, a lot of its power stems from how good the underlying model is. Powerful computers capable of trying out various algorithms to match the sample facts has made it possible to extract models that closely simulate the business goals. The process works iteratively on samples of representative data until the results closely approximate the available factual data.
While data mining is potentially very helpful, its user needs to apply the results in an intelligent way to reap the desired benefit. Not all predictions made make business sense, and can actually be a wasteful exercise in many cases. Used intelligently, it can detect (and predict) fraudulent transactions, defects in an assembly lines and spread of epidemics.
e2e Analytix has executed data mining projects in high tech manufacturing and judicial arbitration domains with success. Clients have benefited immensely by realizing what is causing their product recalls and which panelists are the most productive, thus improving their bottom line and quality of their products |
