Confusion Matrix
Counts how reference classes and mapped classes intersect.
N=[nij]Rows represent reference classes and columns mapped classes (or as displayed).
Best used when: diagnosing which classes are confused.
Classification Accuracy Assessment
Compare a classified raster with independent reference classes and calculate a confusion matrix, Overall Accuracy, Kappa, Producer Accuracy and User Accuracy.
Remote Sensing CoreEnd-User DocumentationTheory + Formula + Parameters1. What This Feature Does
Compare a classified raster with independent reference classes and calculate a confusion matrix, Overall Accuracy, Kappa, Producer Accuracy and User Accuracy.
Cloud workflow: choose data → configure the scientific method → run → review the resulting layer/report. No programming is required.
2. Recommended Workflow
Choose the classified raster and class band.
↓
Choose independent reference points/polygons and the reference class field.
↓
Run the comparison.
↓
Inspect the confusion matrix and class-specific accuracies.
↓
Report sample design and number of validation observations together with the metrics.
3. Theory, Methods & Equations
Overall Accuracy (OA)
Fraction of all evaluated samples/pixels that are correctly classified.
OA = Σnii / N
Best used when: a single overall correctness summary is needed.
Producer Accuracy
Probability that a reference class is correctly mapped; related to omission error.
PAi=nii/ni+
Best used when: assessing omission of a reference class.
User Accuracy
Reliability of a mapped class; related to commission error.
UAi=nii/n+i
Best used when: assessing commission into a mapped class.
Cohen's Kappa
Compares observed agreement with agreement expected from the class marginals.
κ=(po−pe)/(1−pe)
Best used when: a chance-corrected agreement statistic is required, while remembering its known dependence on prevalence/marginals.
4. Input Data
| Input | Type | Requirement | Description |
|---|---|---|---|
| Classified Raster | raster | Required | |
| Reference / Validation Layer | vector | Required |
5. Parameters Available in the Application
| Parameter | Default | Choices / Range | How to Use It |
|---|---|---|---|
| Classification Band | — | — | |
| Reference Class Field | — | — | |
| All Touched Reference Pixels | Off | — |
6. Output & Interpretation
The result should be interpreted according to the selected method and the physical meaning of the input data. Preserve source units, coordinate reference information, NoData meaning, acquisition date, and preprocessing level when comparing results.
Scientific interpretation: an algorithm can produce a numerically valid result even when the input data are unsuitable. Always check masks, units, sensor characteristics, spatial resolution and reference data.
Important: Accuracy metrics are only as defensible as the reference data and sampling design. Validation data should be independent from training data.
7. Best Practices
- Use analysis-ready inputs and remove invalid/cloud/noise artifacts that are not part of the target phenomenon.
- Choose parameters from the physical scale of the data, not only from visual appearance.
- Keep categorical and continuous rasters conceptually separate when selecting interpolation/resampling methods.
- Compare the result with the source image and independent reference information.
- Document the settings used when results will be compared across dates, sensors or study areas.