What is a cluster analysis?
It is an exploratory procedure in which data sets are analysed for possible Similarities in their structures and properties examined are identified. During the analysis, the similarities found in the data sets to be examined are recorded in clusters. For this classification, characteristics such as socio-economic status or the attitude of groups of people are used. For this purpose, a proximity measure is applied, which is a similarity measure.
What is the aim of cluster analysis?
The aim of the cluster analysis is to create a clear Gain structure within large amounts of dataThis is done by arranging them into homogeneous groups, also called clusters. This makes it possible to find people within a data set who have a uniform attitude towards a topic.
What are the methods of cluster analysis?
There are several procedures for such an analysis, within which previously recorded criteria and characteristics are used. Among them are the partitioning and hierarchical procedures. In the partitioning method, the objects to be examined are divided into corresponding clusters by means of previously quantitatively recorded areas. This should contribute to a number of clusters that are as homogeneous as possible.
In contrast, the hierarchical procedure is further subdivided into the divisive and agglomerative procedure. During the divisive procedure, the reference is first directed to a cluster in order to then distribute it to several hierarchically divided clusters. The hierarchical distribution of the clusters takes place in so-called overlapping, non-overlapping and fuzzy groups. During the agglomerative procedure, a summary reduction of several clusters to a manageable number takes place.
An important basic rule that every cluster analysis procedure must observe is that the differences recorded between the individual clusters must be as large as possible and those within a cluster must be as small as possible.
Where is cluster analysis used?
Such an analysis is mostly carried out in market research. It forms the indispensable basis for companies to gain sufficient orientation in their own markets. This makes it possible to divide customers into groups and thus clusters and to conduct target-oriented customer marketing. It guarantees an overview of the decision-making and buying behaviour of customers. This analysis is used for the development of products as well as for the marketing and sales of products. This also gives market research the possibility to divide customers into different buyer types based on their property structures and to adjust their product placements accordingly.
