Comparison of Three Supervised Learning Methods for Digital Soil Mapping: Application to a Complex Terrain in the Ecuadorian Andes

Ließ, Mareike

doi:10.1155/2014/809495

Cited by 21 publications

(15 citation statements)

References 43 publications

(51 reference statements)

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…However, Were et al [87] found SVM as best predictor for the spatial distribution of SOC stock compared to RFR. Rossel et al [88] reported RFR as having better prediction accuracy compared to SGB, while Hitziger et al [89] found the latter superior to the former in soil property prediction. Similarly, SVM and SGB occasionally outperformed RFR in this study.…”

Section: Resultsmentioning

confidence: 99%

High Resolution Mapping of Soil Properties Using Remote Sensing Variables in South-Western Burkina Faso: A Comparison of Machine Learning and Multiple Linear Regression Models

et al. 2017

View full text Add to dashboard Cite

Accurate and detailed spatial soil information is essential for environmental modelling, risk assessment and decision making. The use of Remote Sensing data as secondary sources of information in digital soil mapping has been found to be cost effective and less time consuming compared to traditional soil mapping approaches. But the potentials of Remote Sensing data in improving knowledge of local scale soil information in West Africa have not been fully explored. This study investigated the use of high spatial resolution satellite data (RapidEye and Landsat), terrain/climatic data and laboratory analysed soil samples to map the spatial distribution of six soil properties–sand, silt, clay, cation exchange capacity (CEC), soil organic carbon (SOC) and nitrogen–in a 580 km2 agricultural watershed in south-western Burkina Faso. Four statistical prediction models–multiple linear regression (MLR), random forest regression (RFR), support vector machine (SVM), stochastic gradient boosting (SGB)–were tested and compared. Internal validation was conducted by cross validation while the predictions were validated against an independent set of soil samples considering the modelling area and an extrapolation area. Model performance statistics revealed that the machine learning techniques performed marginally better than the MLR, with the RFR providing in most cases the highest accuracy. The inability of MLR to handle non-linear relationships between dependent and independent variables was found to be a limitation in accurately predicting soil properties at unsampled locations. Satellite data acquired during ploughing or early crop development stages (e.g. May, June) were found to be the most important spectral predictors while elevation, temperature and precipitation came up as prominent terrain/climatic variables in predicting soil properties. The results further showed that shortwave infrared and near infrared channels of Landsat8 as well as soil specific indices of redness, coloration and saturation were prominent predictors in digital soil mapping. Considering the increased availability of freely available Remote Sensing data (e.g. Landsat, SRTM, Sentinels), soil information at local and regional scales in data poor regions such as West Africa can be improved with relatively little financial and human resources.

show abstract

Section: Resultsmentioning

confidence: 99%

High Resolution Mapping of Soil Properties Using Remote Sensing Variables in South-Western Burkina Faso: A Comparison of Machine Learning and Multiple Linear Regression Models

et al. 2017

View full text Add to dashboard Cite

show abstract

“…RF make lots of weak, independent trees, therefore it discerns patterns that otherwise may be disregarded in the cases of few strong trees (Stum et al, 2010). RF is a relatively new method in DSM, it was used to predict topsoil texture classes (Hitziger and Ließ, 2014), soil parent material (Heung et al, 2014), soil organic matter (Wiesmeier et al, 2011) as well as soil types (Brungard et al, 2015;Láng et al, 2016;Stum et al, 2010). Hengl et al (2015) generated numerous soil property predictions (organic carbon, pH, sand, silt and clay fractions, bulk density, cation-exchange capacity, total nitrogen, and exchangeable acidity, Al content and exchangeable bases) of Africa in 250 m resolution by RF.…”

Section: Numerical Classification In Digital Soil Mappingmentioning

confidence: 99%

Compilation of a national soil-type map for Hungary by sequential classification methods

et al. 2018

View full text Add to dashboard Cite

Traditionally in Hungary the soil cover under agricultural and forestry management is typically characterized independently and just approximately identically. Soil data collection is carried out and the databases of soil features are managed irrespectively. As a consequence, nationwide soil maps cannot be considered homogeneously predictive for soils of croplands and forests, plains and hilly/mountainous regions. In order to compile a national soil type map with harmonized legend as well as with spatially relatively homogeneous predictive power and accuracy, the authors unified their resources. Soil profile data originating from the two sources (agriculture and forestry) were cleaned up and harmonized according to a common soil type classification. Various methods were tested for the compilation of the target map: segmentation of a synthesized image consisting of the predictor variables, multi stage classification by Classification and Regression Trees, Random Forests and Artificial Neural Networks. Evaluation of the results showed that the object based, multi-level mapping approach performs significantly better than the simple classification techniques. A combination of best performing classifiers, when each classifier's vote on the same object is weighted according to its confidence in the voted class, led to the final product: a unified, national, soil type map with spatially consistent predictive capabilities.

show abstract

“…According to Hastie et al [ 24 ] random forests do remarkably well with little tuning. In our experience [ 10 , 39 ] tuning random forest and boosted regression tree models can improve prediction results, but tuning needs to be tested carefully for it can also cause overfitting.…”

Section: Introductionmentioning

confidence: 99%

Improving the Spatial Prediction of Soil Organic Carbon Stocks in a Complex Tropical Mountain Landscape by Methodological Specifications in Machine Learning Approaches

Ließ

Schmidt²,

Glaser

2016

PLoS ONE

Self Cite

View full text Add to dashboard Cite

Tropical forests are significant carbon sinks and their soils’ carbon storage potential is immense. However, little is known about the soil organic carbon (SOC) stocks of tropical mountain areas whose complex soil-landscape and difficult accessibility pose a challenge to spatial analysis. The choice of methodology for spatial prediction is of high importance to improve the expected poor model results in case of low predictor-response correlations. Four aspects were considered to improve model performance in predicting SOC stocks of the organic layer of a tropical mountain forest landscape: Different spatial predictor settings, predictor selection strategies, various machine learning algorithms and model tuning. Five machine learning algorithms: random forests, artificial neural networks, multivariate adaptive regression splines, boosted regression trees and support vector machines were trained and tuned to predict SOC stocks from predictors derived from a digital elevation model and satellite image. Topographical predictors were calculated with a GIS search radius of 45 to 615 m. Finally, three predictor selection strategies were applied to the total set of 236 predictors. All machine learning algorithms—including the model tuning and predictor selection—were compared via five repetitions of a tenfold cross-validation. The boosted regression tree algorithm resulted in the overall best model. SOC stocks ranged between 0.2 to 17.7 kg m-2, displaying a huge variability with diffuse insolation and curvatures of different scale guiding the spatial pattern. Predictor selection and model tuning improved the models’ predictive performance in all five machine learning algorithms. The rather low number of selected predictors favours forward compared to backward selection procedures. Choosing predictors due to their indiviual performance was vanquished by the two procedures which accounted for predictor interaction.

show abstract

Comparison of Three Supervised Learning Methods for Digital Soil Mapping: Application to a Complex Terrain in the Ecuadorian Andes

Cited by 21 publications

References 43 publications

High Resolution Mapping of Soil Properties Using Remote Sensing Variables in South-Western Burkina Faso: A Comparison of Machine Learning and Multiple Linear Regression Models

High Resolution Mapping of Soil Properties Using Remote Sensing Variables in South-Western Burkina Faso: A Comparison of Machine Learning and Multiple Linear Regression Models

Compilation of a national soil-type map for Hungary by sequential classification methods

Improving the Spatial Prediction of Soil Organic Carbon Stocks in a Complex Tropical Mountain Landscape by Methodological Specifications in Machine Learning Approaches

Contact Info

Product

Resources

About