Data mining describes the process of extracting data from large sets of information and presenting it in a unique way. This process often is found in business intelligence studies, in which experts mine large sets of data regarding a market or business's operations and attempt to discover previously unrecognized relationships and trends. A data mining model refers to techniques that specialists use to group and present information, as well as the ways in which they can apply information to certain questions and problems.
Many specialists consider data mining regression the most basic and commonly used data mining model. In this process, an expert analyzes a set of data and creates a formula that describes it. Many financial analysts use this technique to make predictions regarding prices and market trends. This model works best in scenarios in which data is expected to remain consistent.
Another popular data mining model is based on association. A specialist can analyze data sets to determine which components often appear together. When two components are paired again and again, a researcher can assume that some association exists between them. For example, a researcher who uses data mining to learn about the performance of a retail store might find that consumers often purchase pens and pencils at the same time that they purchase paper. A manager can use information learned from a data mining model to increase sales by displaying all associated items in once space.
Factor analysis is another common data mining model. In this process, a researcher gathers a number of different variables and attempts to locate factors that determine fluctuations in value. A market researcher, for example, can learn from a customer base how it rates features of similar products. A researcher can then organize this information to illustrate factors that determine consumers' valuation of features. While proponents of this model believe that it can highlight commonality among seemingly disparate variables, some critics believe this model can lead some interpreters to assume causation of certain phenomena when all of the information necessary for determining causation may not be available.
Researchers may use a data mining model based on categorization for simpler problems. Using this technique, specialists organize data by their classifications and tend to organize them in a visual form, such as in a tree or chart. This kind of model is especially helpful in scenarios in which an individual must choose from several options in each category. A designer might find this model useful if in each step of a process he or she can choose from several materials.