Corpus ID: 18397895

Predicting Africa Soil Properties Using Machine Learning Techniques

@inproceedings{Akinola2014PredictingAS,
  title={Predicting Africa Soil Properties Using Machine Learning Techniques},
  author={Iretiayo Akinola and Thomas C Dowd},
  year={2014}
}
Different machine learning algorithms were assessed for estimating five functional soil parameters (SOC content, Calcium content, Phosphorous content, sand content, and pH value). The algorithms used include variants of linear regression and support vector regression. A closer look at the prediction performance for each target revealed that apart from pH, which consistently had worse performance, prediction for the other soil properties was quite satisfactory (RMSE < 0.4). Applying machine… CONTINUE READING

Figures and Tables from this paper

References

SHOWING 1-7 OF 7 REFERENCES
Using data mining to model and interpret soil diffuse reflectance spectra.
  • 626
Predicting soil properties in the tropics
  • 159
LIBSVM: A library for support vector machines
  • 37,343
  • PDF
The Nature of Statistical Learning Theory
  • V. N. Vapnik
  • Mathematics, Political Science
  • Statistics for Engineering and Information Science
  • 2000
  • 36,680