Application of random forest and hierarchical clustering models for crop and fertilizer recommendation to farmers
Uloženo v:
| Název: | Application of random forest and hierarchical clustering models for crop and fertilizer recommendation to farmers |
|---|---|
| Autoři: | Nandom, S.S., Abe, G.T., Gambo, I.P. |
| Zdroj: | Nigerian Journal of Technology; Vol. 44 No. 1 (2025); 114-122 |
| Informace o vydavateli: | African Journals Online (AJOL), 2025. |
| Rok vydání: | 2025 |
| Témata: | Data Analytics, Hierarchical Clustering, Random Forest Algorithm, Fertilizer Recommendation, Crop Recommendation |
| Popis: | Specific recommendations of crop and fertilizer are two critical parts of developing effective agricultural and food policies in Nigeria and other parts of the world. One of the main problems that has negatively affected crop production is the depletion of soil nutrients. Hence maintaining soil nutrients has become a significant concern for farmers. Although fertilizers can be applied manually to increase crop production, it is not optimal since different crops in different fields require different amounts of fertilizer due to soil types, soil fertility levels, and nutrient needs. To effectively and efficiently improve and maintain soil fertility, it is necessary to replace the traditional trial and error method of Nitrogen (N) Potassium (P) and Phosphorus (K) variation at different ratios on untested soils (which most times leads to poor crop yield) with soil testing and fertilizer recommendation using data mining algorithms. This study developed a model to recommend crop and fertilizer using two machine learning algorithms. The RF algorithm, which has shown high level of accuracy in many different agricultural applications, is used for recommending crops, while the hierarchical Clustering algorithm is used for fertilizer recommendation. The models used Crop nutrient requirement and soil sample data for training and testing. The RF and hierarchical algorithm were trained to recommend crop and fertilizer on the basis of multiple biophysical variables and soil nutrients. The system was found effective in recommending crop and fertilizer with an accuracy of 99.70%. The results showed that the model performed effectively and it is versatile machine-learning model for recommending crop and fertilizer due to the high accuracy and precision values. This research pointed out various steps in which a crop and fertilizer recommendation system was achieved using a random forest and hierarchical Clustering algorithms. |
| Druh dokumentu: | Article |
| Popis souboru: | application/pdf |
| ISSN: | 2467-8821 0331-8443 |
| DOI: | 10.4314/njt.v44i1.13 |
| Přístupová URL adresa: | https://www.ajol.info/index.php/njt/article/view/295672 |
| Přístupové číslo: | edsair.doi.dedup.....78bbaf4eeab3bdc3eb8f87434ace6144 |
| Databáze: | OpenAIRE |
Nájsť tento článok vo Web of Science | Abstrakt: | Specific recommendations of crop and fertilizer are two critical parts of developing effective agricultural and food policies in Nigeria and other parts of the world. One of the main problems that has negatively affected crop production is the depletion of soil nutrients. Hence maintaining soil nutrients has become a significant concern for farmers. Although fertilizers can be applied manually to increase crop production, it is not optimal since different crops in different fields require different amounts of fertilizer due to soil types, soil fertility levels, and nutrient needs. To effectively and efficiently improve and maintain soil fertility, it is necessary to replace the traditional trial and error method of Nitrogen (N) Potassium (P) and Phosphorus (K) variation at different ratios on untested soils (which most times leads to poor crop yield) with soil testing and fertilizer recommendation using data mining algorithms. This study developed a model to recommend crop and fertilizer using two machine learning algorithms. The RF algorithm, which has shown high level of accuracy in many different agricultural applications, is used for recommending crops, while the hierarchical Clustering algorithm is used for fertilizer recommendation. The models used Crop nutrient requirement and soil sample data for training and testing. The RF and hierarchical algorithm were trained to recommend crop and fertilizer on the basis of multiple biophysical variables and soil nutrients. The system was found effective in recommending crop and fertilizer with an accuracy of 99.70%. The results showed that the model performed effectively and it is versatile machine-learning model for recommending crop and fertilizer due to the high accuracy and precision values. This research pointed out various steps in which a crop and fertilizer recommendation system was achieved using a random forest and hierarchical Clustering algorithms. |
|---|---|
| ISSN: | 24678821 03318443 |
| DOI: | 10.4314/njt.v44i1.13 |