There are three basic approaches to forming segments:
- Judgment. For example, deciding to segment based on age. This is sometimes known as a priori segmentation. Within Displayr, judgment-based segmentations are usually created by first creating tables and visualizations to explore the data, and then creating new variables.
- Cluster-based segmentation, which involves the use of techniques like latent class analysis, k-means cluster analysis, hierarchical cluster analysis and self-organizing maps to identify clusters of people that are broadly similar (see The Relationship Between Cluster Analysis, Latent Class Analysis and Self-Organizing Maps for an overview of the different algorithms). In Displayr, the options for creating segments include:
- Latent class analysis, via Anything > Advanced Analysis > Cluster > Latent Class Analysis.
- K-Means cluster analysis, via Anything > Advanced Analysis > Cluster > K-Means Cluster Analysis.
- Many other options are available using R, via Calculation > Custom Code.
- Predictive models, designed to identify segments of people (e.g., a CART may be used to identify the people most likely to switch telecommunications provider). These are available in Displayr via Anything > Advanced Analysis > Machine Learning.
Pages in category ‘Segmentation’
The following 11 pages are in this category, out of 11 total.