# kriging

> method of interpolation based on Gaussian process governed by prior covariances

**Wikidata**: [Q225926](https://www.wikidata.org/wiki/Q225926)  
**Wikipedia**: [English](https://en.wikipedia.org/wiki/Kriging)  
**Source**: https://4ort.xyz/entity/kriging

## Summary
Kriging is a method of interpolation based on Gaussian processes governed by prior covariances. It models spatially correlated data to predict values at unobserved locations while quantifying the uncertainty of those predictions.

## Key Facts
- Kriging is named after Danie G. Krige, a South African mining engineer.
- It is classified as an instance of multivariate interpolation, linear regression, and a method.
- The core concept relies on Gaussian processes defined by prior covariance structures.
- Kriging is also known as Gaussian process regression.
- It has a GND identifier: 4339232-5.
- Kriging has an instance relationship with the broader concept of linear regression.
- It belongs to the class of multivariate interpolation methods.

## FAQs
### Q: What is the primary purpose of kriging?
A: Kriging is a statistical interpolation technique used to predict unknown values at specific locations based on spatially dependent data points, while also providing a measure of the prediction's uncertainty.

### Q: How does kriging differ from standard interpolation methods?
A: Unlike standard methods that often only provide a single predicted value, kriging incorporates a Gaussian process model to provide not just the predicted mean but also the variance or credible interval of that prediction at each location.

### Q: Who developed kriging and why is it named after them?
A: Kriging is named after Danie G. Krige, whose early work in mining geostatistics laid the foundation. The method formalizes his approaches to ore reserve estimation using spatial correlation principles.

### Q: What field is kriging primarily used in?
A: Kriging originated in mining and geostatistics but is now widely applied in environmental science, hydrology, geology, computer experiments, and any field requiring spatial prediction of correlated phenomena.

### Q: Is kriging related to machine learning techniques?
A: Yes, kriging is fundamentally a form of Gaussian process regression, placing it within the realm of Bayesian machine learning and spatial statistics, where it serves as a core predictive modeling tool.

## Why It Matters
Kriging matters because it provides a rigorous statistical framework for spatial prediction that explicitly accounts for spatial autocorrelation and quantifies prediction uncertainty. This allows practitioners to not only estimate values at unsampled locations but also understand the confidence in those estimates, which is crucial for decision-making in resource exploration, environmental monitoring, and risk assessment. By leveraging the properties of Gaussian processes, kriging offers a more statistically robust and informative alternative to simple interpolation or deterministic mapping techniques, fundamentally advancing how spatial data is analyzed and used across numerous scientific and engineering disciplines.

## Notable For
- Being explicitly founded on Gaussian processes governed by prior covariances, defining its core statistical nature.
- Its attribution to Danie G. Krige, linking the method directly to its origins in mining geostatistics.
- Its classification simultaneously as a method, an instance of multivariate interpolation, and an instance of linear regression.
- Providing both predicted values and credible intervals (variance) as outputs, embedding uncertainty quantification within its core functionality.

## Body
### Technical Definition
Kriging is a method of interpolation based on Gaussian processes governed by prior covariances. It operates by modeling the spatial correlation between known data points to predict values at unknown locations. A key characteristic is that it outputs not just a single predicted value but also the variance or credible interval associated with that prediction, reflecting the model's uncertainty at the target location.

### Origins & Naming
The method is named after Danie G. Krige, a South African mining engineer. This naming attribution directly links the statistical method to its practical origins in ore reserve estimation and mining geostatistics.

### Classifications
Kriging holds multiple classifications based on its conceptual nature and relationships:
- Instance of: multivariate interpolation
- Instance of: linear regression
- Instance of: method

### Related Entities
- It has a direct relationship with the class `linear regression` (sitelink count: 43), indicating its foundational statistical approach.
- It is a subclass of the broader concept `Gaussian process`.

### Identifiers
- GND ID: 4339232-5
- YSO ID: 3126 (with qualifiers listing its names in Finnish, Swedish, and English: `kriging-menetelmä`, `Krigingmetoden`, `kriging method`)
- Agrovoc ID: c_0f824b4f
- USGS Thesaurus ID: 616
- Freebase ID: /m/02f64l
- Microsoft Academic ID (discontinued): 81692654
- JSTOR Topic ID (archived): kriging
- Commons Category: Kriging
- Stack Exchange Tag: kriging
- GitHub Topic: kriging
- Google Play Store ID: kriging
- Quora Topic: Kriging

### Representation & Dissemination
- An illustrative image is available: [Example of kriging interpolation in 1D](https://commons.wikimedia.org/wiki/Special:FilePath/Example_of_kriging_interpolation_in_1D.png) (describing squares as data points, red curve as kriging interpolation, grey areas as credible intervals, dashed curve as spline).
- An explanatory video is available: [Gaussian Process Animation](https://commons.wikimedia.org/wiki/Special:FilePath/Gaussianprocess.gif).
- The entity has 23 Wikipedia sitelinks across languages: ca, commons, cs, de, en, es, eu, fi, fr, hr.
- The Wikipedia title is "Kriging".

## References

1. Freebase Data Dumps. 2013
2. YSO-Wikidata mapping project
3. Quora
4. [kriging · GitHub Topics](https://github.com/topics/kriging)
5. [OpenAlex](https://docs.openalex.org/download-snapshot/snapshot-data-format)