Segment Analysis
Context
- The term segment typically refers to a subset of similar items within a larger set of observations. For instance, a specific age cohort could be considered a segment within a population. Segments can also refer to subsets of similar products within an industry, e.g., full-size pickup trucks or compact SUVs.
- The similarity of items within a segment can be defined by a single measure, such as age, or by multiple measures combined, such as socioeconomic strata.
- A common research objective is to search for criteria that can identify clusters within a population in order to define new segments. For instance, you can use BayesiaLab's Data Clustering for that purpose.
- Especially for segments generated by clustering algorithms, the identified differences in their high-dimensional characteristics may not always be obvious.
- As opposed to generating new segments, BayesiaLab's Segment Analysis functions are meant to help you understand existing segments, which are identified by the so-called Breakout Variable.
- The Breakout Variable refers to a node that has as many states as there are segments, e.g., Segment A, Segment B, and Segment C.
- Sometimes the segment definitions per se are clear, e.g., Passenger Cars versus Light Trucks, but the characteristics of an associated population, i.e., buyers of Passenger Cars versus buyers of Light Trucks, are not.
- By using the segment name as the Breakout Variable, BayesiaLab can "break out" the corresponding subsets and compare them with each other or compare one particular segment to the overall population.