Scholarlink

<p>Recommending suitable crops considering the soil and climatic-related parameters becomes a crucial concern for sustainable agriculture. Machine learning (ML) techniques can effectively help to recommend suitable crops. However, the accuracy of the recommendations may be limited by the quality and availability of the input data, as certain environmental variables might not be comprehensively covered in all datasets. Therefore, this research proposed Attentive Dilated CNN with Voting Ensemble(ADCVE) to recommend the best suitable crops with maximum accuracy and also developed a predictive mobile application based on the proposed ADCVE model. At first, a total of seven traditional classification algorithms were explored on a dataset generated using Conditional Generative Adversarial Network (CTGAN). Then, the performance of the classification algorithms was analysed independently. Afterwards, the algorithms having an acceptable level of accuracy were integrated with an attentive dilated cnn for highly relevant feature extraction with best ensemble voting classifier that includes Random Forest (RF), Naive Bayes (NB), Decision Tree (DT), Support Vector Machine (SVM), and Logistic Regression (LR). The evaluation performance of the proposed model showed up to 99.12% accuracy. Next, the proposed model was also evaluated on three distinct crop recommendation datasets, and the highest accuracy rates of 99.03%, 98.40%, and 99.12% were observed, respectively. Finally, a user-friendly multilingual crop recommendation mobile application (GoHarvest) was developed to recommend crops accurately.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

GoHarvest: an intelligent crop recommendation system using explainable attentive dilated CNN with voting ensemble(ADCVE) integrating CTGAN method at user end

  • 期刊论文

  • 发表日期：

    2026年3月5日

查看全文
  • Nazmun Nahar Khanom,
  • Md Shofiqul Islam,
  • Jarin Tabassum,
  • Md. Fakhrul Alam Siddiqei,
  • Muhammad Nazrul Islam

摘要

Recommending suitable crops considering the soil and climatic-related parameters becomes a crucial concern for sustainable agriculture. Machine learning (ML) techniques can effectively help to recommend suitable crops. However, the accuracy of the recommendations may be limited by the quality and availability of the input data, as certain environmental variables might not be comprehensively covered in all datasets. Therefore, this research proposed Attentive Dilated CNN with Voting Ensemble(ADCVE) to recommend the best suitable crops with maximum accuracy and also developed a predictive mobile application based on the proposed ADCVE model. At first, a total of seven traditional classification algorithms were explored on a dataset generated using Conditional Generative Adversarial Network (CTGAN). Then, the performance of the classification algorithms was analysed independently. Afterwards, the algorithms having an acceptable level of accuracy were integrated with an attentive dilated cnn for highly relevant feature extraction with best ensemble voting classifier that includes Random Forest (RF), Naive Bayes (NB), Decision Tree (DT), Support Vector Machine (SVM), and Logistic Regression (LR). The evaluation performance of the proposed model showed up to 99.12% accuracy. Next, the proposed model was also evaluated on three distinct crop recommendation datasets, and the highest accuracy rates of 99.03%, 98.40%, and 99.12% were observed, respectively. Finally, a user-friendly multilingual crop recommendation mobile application (GoHarvest) was developed to recommend crops accurately.