Application of machine learning methods to spatial interpolation of environmental variables

Li, Jin; Heap, Andrew D.; Potter, A. W.; Daniell, James

doi:10.1016/j.envsoft.2011.07.004

Cited by 312 publications

(216 citation statements)

References 30 publications

Supporting

Mentioning

196

Contrasting

Unclassified

Order By: Relevance

“…This concept could readily be adapted to a variety of other situations where local physical properties must be estimated from remotely sensed data. Further discussion of the use of random forests for interpola- tion between samples can be found in, for example, Li et al (2011), who test a variety of techniques including random forests and support vector machines in order to produce a regional map of mud content in sediments based on discrete sampling. Another example comes from Bhattacharya et al (2007), who construct models of sediment transport using a variant of decision trees where each "leaf" is a linear regression function, rather than a single classification; the decision tree is essentially used to choose which mathematical model to apply in each particular combination of circumstances.…”

Section: Decision Trees and Random Forestsmentioning

confidence: 99%

“…In geomorphology, a common application of machine learning for regression and interpolation is to link widely available remote sensing measurements with underlying parameters of interest that cannot be measured directly: for example, sediment and chlorophyll content of water from colour measurements (Krasnopolsky and Schiller, 2003) or marine sediment properties from bathymetry and proximity to the coast (Li et al, 2011;Martin et al, 2015).…”

Section: Regression and Interpolationmentioning

confidence: 99%

See 1 more Smart Citation

An introduction to learning algorithms and potential applications in geomorphometry and Earth surface dynamics

Valentine

Kalnins

2016

Earth Surf. Dynam.

View full text Add to dashboard Cite

Abstract. "Learning algorithms" are a class of computational tool designed to infer information from a data set, and then apply that information predictively. They are particularly well suited to complex pattern recognition, or to situations where a mathematical relationship needs to be modelled but where the underlying processes are not well understood, are too expensive to compute, or where signals are over-printed by other effects. If a representative set of examples of the relationship can be constructed, a learning algorithm can assimilate its behaviour, and may then serve as an efficient, approximate computational implementation thereof. A wide range of applications in geomorphometry and Earth surface dynamics may be envisaged, ranging from classification of landforms through to prediction of erosion characteristics given input forces. Here, we provide a practical overview of the various approaches that lie within this general framework, review existing uses in geomorphology and related applications, and discuss some of the factors that determine whether a learning algorithm approach is suited to any given problem.

show abstract

Section: Decision Trees and Random Forestsmentioning

confidence: 99%

Section: Regression and Interpolationmentioning

confidence: 99%

An introduction to learning algorithms and potential applications in geomorphometry and Earth surface dynamics

Valentine

Kalnins

2016

Earth Surf. Dynam.

View full text Add to dashboard Cite

show abstract

“…Previous work by Shi et al [1] demonstrated the effectiveness of incorporating land use type and soil type to improve interpolation simulation of soil properties. In addition, many studies have identified topography as an important auxiliary element [3,19], but previous research results suggest it is not a key factor in the study area [10]. Therefore, integration of secondary variables in this study should have an important influence on interpolation accuracy.…”

Section: Introductionmentioning

confidence: 99%

“…Spatial interpolation is the main method used to evaluate continuous changes in soil properties [1], as well as being an important research tool in the fields of 'digital soil' and 'pedometrics' mapping [2]. Current spatial interpolation methods mainly originate from discrete modern mathematical theories (function theory and differential geometry), and can be largely classified into three groups [3]: (1) deterministic or non-geostatistical methods (e.g., inverse distance weighting, IDW), (2) geostatistical methods (e.g., ordinary kriging, OK), and (3) combined methods (e.g., regression kriging). These methods are often data-or even variable-specific and their performance depends on many factors.…”

Section: Introductionmentioning

confidence: 99%

Adaptive Surface Modeling of Soil Properties in Complex Landforms

Liu

Zhang

Yan

et al. 2017

IJGI

View full text Add to dashboard Cite

Spatial discontinuity often causes poor accuracy when a single model is used for the surface modeling of soil properties in complex geomorphic areas. Here we present a method for adaptive surface modeling of combined secondary variables to improve prediction accuracy during the interpolation of soil properties (ASM-SP). Using various secondary variables and multiple base interpolation models, ASM-SP was used to interpolate soil K + in a typical complex geomorphic area (Qinghai Lake Basin, China). Five methods, including inverse distance weighting (IDW), ordinary kriging (OK), and OK combined with different secondary variables (e.g., OK-Landuse, OK-Geology, and OK-Soil), were used to validate the proposed method. The mean error (ME), mean absolute error (MAE), root mean square error (RMSE), mean relative error (MRE), and accuracy (AC) were used as evaluation indicators. Results showed that: (1) The OK interpolation result is spatially smooth and has a weak bull's-eye effect, and the IDW has a stronger 'bull's-eye' effect, relatively. They both have obvious deficiencies in depicting spatial variability of soil K + . (2) The methods incorporating combinations of different secondary variables (e.g., ASM-SP, OK-Landuse, OK-Geology, and OK-Soil) were associated with lower estimation bias. Compared with IDW, OK, OK-Landuse, OK-Geology, and OK-Soil, the accuracy of ASM-SP increased by 13.63%, 10.85%, 9.98%, 8.32%, and 7.66%, respectively. Furthermore, ASM-SP was more stable, with lower MEs, MAEs, RMSEs, and MREs. (3) ASM-SP presents more details than others in the abrupt boundary, which can render the result consistent with the true secondary variables. In conclusion, ASM-SP can not only consider the nonlinear relationship between secondary variables and soil properties, but can also adaptively combine the advantages of multiple models, which contributes to making the spatial interpolation of soil K + more reasonable.

show abstract

“…Third, a new hybrid interpolation method is proposed by combining the SVM with inverse distance weighting, ordinary kriging and other local interpolation methods. used SVM, ordinary kriging, inverse distance squared and their combinations to conduct spatial interpolation of mud content samples in the southwest Australian margin; however, the SVM and its combination with ordinary kriging or inverse distance squared are not suitable for the spatial interpolation effect of mud content in this region [16]. The interpolation precision of SVM is affected by normalization range, the selection of kernel functions and the reasonable setting of the insensitive loss parameter and penalty parameter, etc.…”

Section: Introductionmentioning

confidence: 99%

Application of a Hybrid Interpolation Method Based on Support Vector Machine in the Precipitation Spatial Interpolation of Basins

Zhang

Liu

Wang³

et al. 2017

Water

View full text Add to dashboard Cite

Abstract:In this paper, we applied the support vector machine (SVM) to the spatial interpolation of the multi-year average annual precipitation in the Three Gorges Region basin. By combining it with the inverse distance weighting and ordinary kriging method, we constructed the SVM residual inverse distance weighting, as well as the SVM residual kriging precipitation interpolation model and compared them with the inverse distance weighting, ordinary kriging, linear regression residual inverse distance weighting and linear regression residual kriging interpolation methods. The TRMM 3B43 V7 satellite precipitation information, which is processed by the latest revision algorithm, is used as the auxiliary variable for ground site precipitation interpolation along with latitude and elevation. Our results show that: (1) adding the TRMM 3B43 V7 satellite precipitation data as an auxiliary variable significantly improves the interpolation accuracy of the linear regression equation and SVM model; (2) the support vector machine hybrid interpolation method obtains superior interpolation results compared to the inverse distance weighting method, ordinary kriging method and linear regression hybrid interpolation method; (3) the interpolation accuracy of the SVM hybrid interpolation method depends on the SVM fitting degree, so we should choose a suitable fitting accuracy rather than the highest fitting accuracy; (4) the linear regression equation has a greater degree of dependency on the TRMM data than the SVM. The SVM accepts the TRMM data information while better maintaining its independence, taking into account that the TRMM data linear regression and linear regression hybrid interpolation method are not suitable for TRMM data evaluation.

show abstract

Application of machine learning methods to spatial interpolation of environmental variables

Cited by 312 publications

References 30 publications

An introduction to learning algorithms and potential applications in geomorphometry and Earth surface dynamics

An introduction to learning algorithms and potential applications in geomorphometry and Earth surface dynamics

Adaptive Surface Modeling of Soil Properties in Complex Landforms

Application of a Hybrid Interpolation Method Based on Support Vector Machine in the Precipitation Spatial Interpolation of Basins

Contact Info

Product

Resources

About