HE1-20 Trend and time-series decomposition Also serves: Project forward.
Separates a series into trend, seasonal and irregular parts and produces plots and a fitted trend.
What it does
Separates trend, seasonal and irregular parts of a series. Artefact: plots and a fitted trend; read breaks and seasonality.
When to use it
Use when you need to know whether a rising or falling series reflects a trend, a seasonal pattern or noise.
When it does not fit
Definitions changed during the series.
Result format
The result has plots of the series and its separated trend, seasonal and irregular parts, with a fitted trend. It names the breaks in the series and the seasonal pattern.
Requirements on the material
| Property | Requirement type | Requirement |
|---|---|---|
| Structure | Applies in some cases | requires=structured |
| Content type | Always applies | requires=time series |
| Unit of analysis | Always applies | requires=variable; requires=dataset |
| Variable types | Raises a warning | prefers=interval; prefers=ratio; prefers=counts or fractions |
| Time structure | Applies in some cases | requires=aggregate time series |
| Level of detail available | Always applies | requires=aggregate tables |
| Appraises source credibility | Applies in some cases | no |
Not restricted on: How the sample was chosen, Number of sources, Whether sources can be joined, Amount of material, Known data problems, Producer interest, Source dependence, Origin, Document genre, Language, Personal information, Security classification, Licence.
| Minimum material | at least two full seasonal cycles for seasonal decomposition |
|---|---|
| Additional sources needed | none |
| Not for (rules) | none |
| Not for (text) | Definitions or measurement changed during the series, which produces false breaks. |
Choosing and ordering
| Kind of technique | Primary technique, runs in step 5.3 |
|---|---|
| Phase tag | analyse |
| Other phase tags | none |
| Kind of finding | trend; pattern |
| Question answered | what; how much |
| Strongest statement it supports | describe |
| Needs as input |
|
| Makes as output |
|
| Must exist first | None |
| Loop target | proceed with limit stated |
Library checks by stakes
Each stakes level includes the checks of the level below it. Checks that need a person apply at high stakes only, so the process leaves them out at lower stakes (decision 38). Repeatability: deterministic.
| Stakes | Check | Person involved | Counts as a different method |
|---|---|---|---|
| exploratory | input validation | none | no |
| decision-informing | input validation | none | no |
| decision-informing | spot check | optional | no |
| high-stakes | input validation | none | no |
| high-stakes | spot check | optional | no |
| high-stakes | independent review | needed | no |
| high-stakes | provenance log | none | no |
Failure modes and limits
| Model suitability | model plus code |
|---|---|
| Typical model failure | Fits a decomposition by eye or in prose without running it, and reads a definition change as a trend break. |
| Error sources | input data; statistical assumptions; data transformation |
| Certainty limits | Certainty drops with short series, definition breaks and outliers (imprecision and risk of bias). |
Related techniques
- should follow PE-04 Data profiling and rule-based validation (data audit; data quality assessment; edit checks): profile the series for gaps and definition changes first
- pairs with PJ-01 Trend extrapolation and time-series forecasting (exponential smoothing, autoregressive integrated moving average models, trend impact analysis): decomposed components inform forecasting
Evidence and tag basis
Unconfirmed. A forecasting textbook was listed and not opened.
Tag basis: input profile inference; selection inference; sequencing inference; robustness inference. All tag values are inference unless a source is named.