Evaluation of the influence of soil parameters on vegetation growth using machine learning approaches
| dc.contributor.advisor | Fehér, Zsolt Zoltán | |
| dc.contributor.advisor | Nagy, Attila | |
| dc.contributor.author | Bimantara, Dhimas Sigit | |
| dc.contributor.department | DE--Mezőgazdaság- Élelmiszertudományi és Környezetgazdálkodási Kar | |
| dc.date.accessioned | 2023-05-04T09:42:04Z | |
| dc.date.available | 2023-05-04T09:42:04Z | |
| dc.date.created | 2023-05-03 | |
| dc.description.abstract | The impact of climate change on crop production has led to increased adoption of precision farming, which combines advanced technologies such as sensors, information systems, and enhanced machinery. Machine learning algorithms have become an essential tool in precision agriculture, helping analysis tasks to predict and optimize crop yields. The study collected data from the Nyirbator research station in Hungary, and Sentinel-2 data was used to obtain the NDVI values of maize at different phenological phases. the analysis is performed using ArcGIS ML and Python. The study also employs the Shapley Additive Explanation technique to interpret the model's predictions and identify the soil elements that contribute the most to the prediction of the NDVI value. Random Forest Regression is found to be the most suitable ML algorithm for predicting the NDVI of maize. The Random Forest algorithm achieves the lowest test RMSE in Python analysis, indicating that it is effective at capturing the complex relationship between soil parameters and NDVI values. Shapley values were used to break down the predictions of a black-box model and identify the soil elements that contribute the most to the prediction of the NDVI value, providing insights into the model's interpretation. The Random Forest model developed in this study provides a promising method for predicting the NDVI value of maize at different stages of growth. It is comparable to previous research in agricultural applications using Random Forest models, which could help researchers and practitioners select appropriate machine-learning algorithms for this task. The model is determined by the problem's specific requirements and constraints, making it a viable option for improving farm operations' efficiency. | |
| dc.description.corrector | KE | |
| dc.description.course | Agricultural Environmental Management Engineering | |
| dc.description.degree | MSc/MA | |
| dc.format.extent | 81 | |
| dc.identifier.uri | https://hdl.handle.net/2437/351818 | |
| dc.language.iso | en | |
| dc.rights.access | Hozzáférhető a 2022 decemberi felsőoktatási törvénymódosítás értelmében. | |
| dc.subject | machine learning | |
| dc.subject | random forest | |
| dc.subject | soil | |
| dc.subject | maize | |
| dc.subject | precision agriculture | |
| dc.subject.dspace | DEENK Témalista::Mezőgazdaságtudomány | |
| dc.title | Evaluation of the influence of soil parameters on vegetation growth using machine learning approaches |