Models for feature selection and efficient crop yield prediction in the groundnut production
Tamil Nadu ranks high in groundnut production in India. The yield prediction of the crop over Tamil Nadu will be highly useful in improving the efficiency of the production. This article aims to identify an efficient machine learning model to predict the groundnut crop yield and analyse the performa...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Czech Academy of Agricultural Sciences
2022-03-01
|
| Series: | Research in Agricultural Engineering |
| Subjects: | |
| Online Access: | https://rae.agriculturejournals.cz/artkey/rae-202203-0003_models-for-feature-selection-and-efficient-crop-yield-prediction-in-the-groundnut-production.php |
| Summary: | Tamil Nadu ranks high in groundnut production in India. The yield prediction of the crop over Tamil Nadu will be highly useful in improving the efficiency of the production. This article aims to identify an efficient machine learning model to predict the groundnut crop yield and analyse the performance of the tested models. The study used the irrigation, rainfall, area and production data as factors for the groundnut crop yield across the districts of Tamil Nadu. This article identified the best set of features for training the models and studied various prediction models to evaluate the performance on the collected data. The trained and tested data were evaluated using various performance measures. The results of the study show that LASSO and ElasticNet provide the optimal results with the lowest RMSE and RRMSE values of 491.603 and 490.931 kg.ha-1, 20.68 and 20.66%, respectively. The models showed the lowest MAE and RMAE values as well (333.154 and 331.827 kg.ha-1 and 14.53%, 14.51%, respectively) when compared to other models. The identification of the right time to sow and area to irrigate through feature selection and the prediction of the yield will improve the yield of the groundnut crops. This helps farmers to make practical decisions and reap the benefits. |
|---|---|
| ISSN: | 1212-9151 1805-9376 |