Enhancing Crop Recommendation with SMOTE-Augmented Machine Learning Algorithms

Gözde Alp; Fatih Soygazi

Article Details

Enhancing Crop Recommendation with SMOTE-Augmented Machine Learning Algorithms

Authors:

Gözde Alp | Fatih Soygazi

Publisher:

Murat YAKAR

Main Topics:

Computer Software | Software Engineering (Other)

Journal:

Turkish Journal of Engineering Vol.9 Issue 4, pp.612-620

Document Type:

Article

Publication Date:

8/2025

Document Language:

English

Document ID (Org No.):

2001360051

pISSN:

2587-1366

Keywords:

Abstract:

 AAs the agricultural landscape continues to evolve, the synergy between the crop recommendation and machine learning ML innovations holds the potential to revolutionize the way farmers make decisions, ultimately driving sustainable and efficient food production. This study investigates the efficacy of various machine learning algorithms for crop recommendation, a crucial aspect of precision agriculture. Six prominent algorithms are evaluated utilizing the Crop Recommendation Dataset, those are: Attentive Interpretable Tabular Learning, Extreme Gradient Boosting XGBoost, K-Nearest Neighbors, Decision Tree, LightGBM, and Random Forest RF. Synthetic Minority Oversampling Technique SMOTE is applied to each algorithm to increase the number of samples in the dataset. When datasets are small, models may not be able to learn enough features to classify problems effectively in real-time situations. Data augmentation, including techniques like SMOTE, appears as a way to surpass this limitation by increasing the amount of available data. Our results demonstrate the effectiveness of these algorithms in accurately predicting suitable crops based on environmental and soil parameters. Notably, XGBoost achieved an accuracy of 0.9909, while RF combined with SMOTE attained the highest accuracy of 0.9981. This superior performance is attributed to SMOTEs ability to generate a balanced dataset by increasing the number of samples in each class based on the available data, thereby enhancing the models predictive capability for crop recommendation. The key contribution of this study is the demonstration that applying SMOTE for data augmentation can improve the predictive accuracy of various machine learning algorithms. This research underscores the potential of machine learning, particularly ensemble methods coupled with oversampling techniques, in driving data-driven agricultural practices for enhanced crop yield and resource optimization.

 AAs the agricultural landscape continues to evolve, the synergy between the crop recommendation and machine learning ML innovations holds the potential to revolutionize the way farmers make decisions, ultimately driving sustainable and efficient food production. This study investigates the efficacy of various machine learning algorithms for crop recommendation, a crucial aspect of precision agriculture. Six prominent algorithms are evaluated utilizing the Crop Recommendation Dataset, those are: Attentive Interpretable Tabular Learning, Extreme Gradient Boosting XGBoost, K-Nearest Neighbors, Decision Tree, LightGBM, and Random Forest RF. Synthetic Minority Oversampling Technique SMOTE is applied to each algorithm to increase the number of samples in the dataset. When datasets are small, models may not be able to learn enough features to classify problems effectively in real-time situations. Data augmentation, including techniques like SMOTE, appears as a way to surpass this limitation by increasing the amount of available data. Our results demonstrate the effectiveness of these algorithms in accurately predicting suitable crops based on environmental and soil parameters. Notably, XGBoost achieved an accuracy of 0.9909, while RF combined with SMOTE attained the highest accuracy of 0.9981. This superior performance is attributed to SMOTEs ability to generate a balanced dataset by increasing the number of samples in each class based on the available data, thereby enhancing the models predictive capability for crop recommendation. The key contribution of this study is the demonstration that applying SMOTE for data augmentation can improve the predictive accuracy of various machine learning algorithms. This research underscores the potential of machine learning, particularly ensemble methods coupled with oversampling techniques, in driving data-driven agricultural practices for enhanced crop yield and resource optimization.

 AAs the agricultural landscape continues to evolve, the synergy between the crop recommendation and machine learning ML innovations holds the potential to revolutionize the way farmers make decisions, ultimately driving sustainable and efficient food production. This study investigates the efficacy of various machine learning algorithms for crop recommendation, a crucial aspect of precision agriculture. Six prominent algorithms are evaluated utilizing the Crop Recommendation Dataset, those are: Attentive Interpretable Tabular Learning, Extreme Gradient Boosting XGBoost, K-Nearest Neighbors, Decision Tree, LightGBM, and Random Forest RF. Synthetic Minority Oversampling Technique SMOTE is applied to each algorithm to increase the number of samples in the dataset. When datasets are small, models may not be able to learn enough features to classify problems effectively in real-time situations. Data augmentation, including techniques like SMOTE, appears as a way to surpass this limitation by increasing the amount of available data. Our results demonstrate the effectiveness of these algorithms in accurately predicting suitable crops based on environmental and soil parameters. Notably, XGBoost achieved an accuracy of 0.9909, while RF combined with SMOTE attained the highest accuracy of 0.9981. This superior performance is attributed to SMOTEs ability to generate a balanced dataset by increasing the number of samples in each class based on the available data, thereby enhancing the models predictive capability for crop recommendation. The key contribution of this study is the demonstration that applying SMOTE for data augmentation can improve the predictive accuracy of various machine learning algorithms. This research underscores the potential of machine learning, particularly ensemble methods coupled with oversampling techniques, in driving data-driven agricultural practices for enhanced crop yield and resource optimization.

 AAs the agricultural landscape continues to evolve, the synergy between the crop recommendation and machine learning ML innovations holds the potential to revolutionize the way farmers make decisions, ultimately driving sustainable and efficient food production. This study investigates the efficacy of various machine learning algorithms for crop recommendation, a crucial aspect of precision agriculture. Six prominent algorithms are evaluated utilizing the Crop Recommendation Dataset, those are: Attentive Interpretable Tabular Learning, Extreme Gradient Boosting XGBoost, K-Nearest Neighbors, Decision Tree, LightGBM, and Random Forest RF. Synthetic Minority Oversampling Technique SMOTE is applied to each algorithm to increase the number of samples in the dataset. When datasets are small, models may not be able to learn enough features to classify problems effectively in real-time situations. Data augmentation, including techniques like SMOTE, appears as a way to surpass this limitation by increasing the amount of available data. Our results demonstrate the effectiveness of these algorithms in accurately predicting suitable crops based on environmental and soil parameters. Notably, XGBoost achieved an accuracy of 0.9909, while RF combined with SMOTE attained the highest accuracy of 0.9981. This superior performance is attributed to SMOTEs ability to generate a balanced dataset by increasing the number of samples in each class based on the available data, thereby enhancing the models predictive capability for crop recommendation. The key contribution of this study is the demonstration that applying SMOTE for data augmentation can improve the predictive accuracy of various machine learning algorithms. This research underscores the potential of machine learning, particularly ensemble methods coupled with oversampling techniques, in driving data-driven agricultural practices for enhanced crop yield and resource optimization.

Article Details

Enhancing Crop Recommendation with SMOTE-Augmented Machine Learning Algorithms

Authors:

Gözde Alp | Fatih Soygazi

Publisher:

Murat YAKAR

Main Topics:

Computer Software | Software Engineering (Other)

Journal:

Turkish Journal of Engineering Vol.9 Issue 4, pp.612-620

Document Type:

Article

Publication Date:

8/2025

Document Language:

English

Document ID (Org No.):

2001360051

pISSN:

2587-1366

Agriculture

Crop Recommendation

Cropping Strategy

Machine Learning

Sampling

SMOTE

Issues

Article Details

Enhancing Crop Recommendation with SMOTE-Augmented Machine Learning Algorithms

Authors: Gözde Alp | Fatih Soygazi

Publisher: Murat YAKAR

Main Topics: Computer Software | Software Engineering (Other)

Journal: Turkish Journal of Engineering Vol.9 Issue 4, pp.612-620

Document Type: Article

Publication Date: 8/2025

Document Language: English

Document ID (Org No.): 2001360051

pISSN: 2587-1366

Agriculture

Crop Recommendation

Cropping Strategy

Machine Learning

Sampling

SMOTE

Issues

Authors:

Gözde Alp | Fatih Soygazi

Publisher:

Murat YAKAR

Main Topics:

Computer Software | Software Engineering (Other)

Journal:

Turkish Journal of Engineering Vol.9 Issue 4, pp.612-620

Document Type:

Article

Publication Date:

8/2025

Document Language:

English

Document ID (Org No.):

2001360051

pISSN:

2587-1366