Selection of Environmental Covariates for Classifier Training Applied in Digital Soil Mapping

Campos, Alcinei Ribeiro; Giasson, Élvio; Costa, José Janderson Ferreira; Machado, Israel Rosa; Silva, Elisângela Benedet da; Bonfatti, Benito Roberto

doi:10.1590/18069657rbcs20170414

Cited by 8 publications

(3 citation statements)

References 28 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The negative correlation works in the same way but with a negative sign. Pearson's correlation is a frequent method for variable selection, which has been used for the prediction of soil classes (Camera et al, 2017;Campos et al, 2018), soil depth (Camera et al, 2017Lu et al, 2019) and soil organic matter (Chen et al, 2022).…”

Section: Correlationmentioning

confidence: 99%

“…Variable selection has been widely used in different fields such as biomedicine and bioinformatics (Guyon et al, 2002;Saeys et al, 2007;Osl et al, 2009) or text classification problems (Forman, 2003). In soil science, the variable selection have been used for the prediction of soil organic carbon (Xiong et al, 2014;Fitzpatrick et al, 2016;Lie et al, 2016;Keskin et al, 2019), soil parent material (Heung et al, 2014), soil organic matter (Chen et al, 2022), soil depth (Tesfa et al, 2009;Camera et al, 2017;Castro Franco et al, 2017;Lu et al, 2019) and soil classes (Behrens et al, 2010;Brungard et al, 2015;Camera et al, 2017;Campos et al, 2018). So far, there are hardly any works where the selection of variables has been applied for the prediction of AS soils.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Improving prediction accuracy for acid sulfate soil mapping by means of variable selection

Estévez

Mattbäck

Boman

et al. 2023

Front. Environ. Sci.

View full text Add to dashboard Cite

Acid sulfate soils can cause environmental damage and geotechnical problems when drained or exposed to oxidizing conditions. This makes them one of the most harmful soils found in nature. In order to reduce possible damage derived from this type of soil, it is fundamental to create occurrence maps showing their localization. Nowadays, occurrence maps can be created using machine learning techniques. The accuracy of these maps depends on two factors: the dataset and the machine learning method. Previously, different machine learning methods were evaluated for acid sulfate soil mapping. To improve the precision of the acid sulfate soil probability maps, in this qualitative modeling study we have added more environmental covariates (17 in total). Since a greater number of covariates does not necessarily imply an improvement in the prediction, we have selected the most relevant environmental covariates for the classification and prediction of acid sulfate soils. For this, we have applied eleven different variable selection methods. The predictive abilities of each group of selected variables have been analyzed using Random Forest and Gradient Boosting. We show that the selection of each environmental covariate as well as the relationship between them are extremely important for an accurate prediction of acid sulfate soils. Among the variable selection methods analyzed, Random Forest stands out, as it is the one that has best selected the relevant covariates for the classification of these soils. Furthermore, the combination of two variable selection methods can improve the prediction of the model. Contrary to the general belief, a low correlation between the covariates does not guarantee a good performance of the model. In general, Random Forest has given better results in the prediction than Gradient Boosting. From the best results obtained, an acid sulfate soils occurrence map has been created. Compared with previous studies in the same area, variable selection has improved the accuracy by 15%–17% for the models based on Random Forest. The present study confirms the importance of variable selection for the prediction of acid sulfate soils.

show abstract

Section: Correlationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

Improving prediction accuracy for acid sulfate soil mapping by means of variable selection

Estévez

Mattbäck

Boman

et al. 2023

Front. Environ. Sci.

View full text Add to dashboard Cite

show abstract

“…Of the 334 studies, 263 presented cartographic scale and spatial resolution (pixel size) information used (Table 3); 38 studies were found incompatible with the planimetric PEC-PCD, since their pixel size is higher than that indicated at the cartographic scale. (Giasson et al, 2006) 2008 (Figueiredo et al, 2008) 2009 (Crivelenti et al, 2009) 2010 (Chagas et al, 2010); (Coelho and Giasson, 2010) (Höfig et al, 2014); (Teske et al, 2014) 2015 (Bagatini et al, 2015); (Giasson et al, 2015); (Teske et al, 2015a); (Teske et al, 2015b); (Vasques et al, 2015) 2016 (Arruda et al, 2016); (Bagatini et al, 2016); (Demattê et al, 2016); (Dias et al, 2016); (Henrique et al, 2016); (Pelegrino et al, 2016) 2017 (Chagas et al, 2017); (Wolski et al, 2017) 2018 (Costa et al, 2018); (Meier et al, 2018) 2019 (Campos et al, 2019a); (Campos et al, 2019b); (Moura-Bueno et al, 2019); (Silva et al, 2019); (Silvero et al, 2019) All learning algorithms were assigned to a type of classifier such as Artificial Neural Network (ANN), Bayes classifiers, Decision Tree (DT), Logistic Regression (LR) and Support Vector Machine (SVM). Approximately 95 % of the studies used DT, ANN and LR classifiers (Table 4).…”

Section: Descriptive Statistics Of the Data Extracted From The Studiesmentioning

confidence: 99%

Digital soil class mapping in Brazil: a systematic review

Coelho

Giasson

Campos

et al. 2021

Sci. agric. (Piracicaba, Braz.)

Self Cite

View full text Add to dashboard Cite

In Brazil several digital soil class mapping studies were carried out from 2006 onwards to maximize the use of existing maps and information and to provide estimates for wider areas. However, there is no consensus on which methods have produced superior results in the predictive value of soil maps. This study conducts a systematic review of digital soil class mapping in Brazil and aims to analyze the factors which can improve the accuracy of digital soil class maps. Data from 334 digital soil class mapping studies were grouped and analyzed by Student's t-test, Wilcoxon-Mann-Whitney test and Kruskal-Wallis test. When conventional maps were used for validation, the studies showed average values of 63 % and when field samples were used, 56 % for Overall Accuracy. Studies compatible with the Planimetric Cartographic Accuracy Standard for Digital Cartographic Products (PEC-PCD) averaged between 4 % and 15 % higher accuracy than those of the incompatible group. There seems to be no evidence that increasing the number of variables and samples results in more accurate soil map prediction, but studies using variables related to four soil-forming factors enhanced accuracy. From a density of 0.08 MU km-2 and upwards, it became more difficult for studies to obtain greater accuracy. Artificial neural network classifiers and Decision Tree models seem to be producing more accurate digital soil class maps.

show abstract

Feature Selection Models for Data Classification: Wrapper Model vs Filter Model

Khtoom

Wedyan

2020

Learning and Analytics in Intelligent Systems

View full text Add to dashboard Cite

Selection of Environmental Covariates for Classifier Training Applied in Digital Soil Mapping

Cited by 8 publications

References 28 publications

Improving prediction accuracy for acid sulfate soil mapping by means of variable selection

Improving prediction accuracy for acid sulfate soil mapping by means of variable selection

Digital soil class mapping in Brazil: a systematic review

Feature Selection Models for Data Classification: Wrapper Model vs Filter Model

Contact Info

Product

Resources

About