Sequence Characteristics Indicators
Sequence characteristics indicators summarize the shape of each individual trajectory. They are useful before clustering, when describing typologies, or when turning sequences into variables for downstream models.
Choose an Indicator Family
| Question | Useful pages |
|---|---|
| How long are the observed trajectories? | Sequence Length |
| How long do states or spells last? | Spell Durations, Mean Spell Duration, Duration Standard Deviation |
| How many states or subsequences appear? | Visited States, Number of Subsequences |
| How complex or unstable is a trajectory? | Within-Sequence Entropy, Complexity Index, Turbulence, Volatility |
| How do states change over time? | Number of Transitions, Cross-Sectional Entropy |
| Do trajectories move toward better or worse states? (requires a ranked state order) | Positive-Negative Indicators, Integration Index, Badness Index, Degradation Index, Precarity Index, Insecurity Index |
| How do I translate familiar TraMineR indicators? | Sequenzo and TraMineR Mapping |
How to Use Indicators
Sequence characteristics indicators can be reported descriptively, plotted by group, used to interpret clusters, or passed into feature-selection and regression workflows. Use distance-based methods when the research question is about the whole sequence structure, and use indicators when a smaller number of interpretable sequence features is enough.