Cluster analysis is one of the methods of analysis companies usually use in their statistical and management techniques. Its main work is to try and reduce a larger set of data to a more meaningful smaller group of objects. The companies usually use cluster analysis which is accomplished on similarity of objects basis across a set of particular characteristics.
When using this technique, analyst have always identified that outliers are a big problem which arise from very many variables which are irrelevant. It is common knowledge among the analysts that the sample used should be able to be a representative of the population. They also advice that it is better to use uncorrelated factors to achieve desirable results. Major clustering methods used by companies include hierarchical, which takes an example of a tree like example and mostly deals with smaller sets of data. Other clustering methods include nonhierarchical which is the opposite of hierarchical and the third one is a combination of both clustering methods. Clusters should always be different, reachable, and measurable and should always be profitable. (William C. Black, 1998)