Scholarlink

<p>Agriculture is one of India's important sectors, impacting GDP and people's lives. This paper focuses on physical, chemical, and biological factors that affect soil fertility since it is the foundation of agricultural yield. However, conventional soil testing techniques are tedious and do not offer any information that can be used in precision farming. To this end, this study aims to use logistic regression, support vector machines, decision trees, random forests, and k-nearest neighbors to predict the soil fertility levels from the critical soil parameters. The key advance is to use a wide set of soil parameters, including macro and micronutrients, and soil physico-chemical properties like pH, OC, and EC to build predictive models. The results show that using the random forest algorithm is more efficient than other models, with an accuracy of 99%, followed by decision trees at 98%. These results show that machine learning can accurately predict soil fertility while minimizing costs. The study mentioned above aims to help farmers make informed decisions and enhance crop production by establishing the most effective predictive model for identifying potential soil conditions. The authors provide a set of recommendations to incorporate the findings and implications of the study into machine learning applications for improving agricultural yields and efficiency.</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

Predictive analysis of soil fertility using supervised machine learning techniques for enhanced agricultural productivity

  • 期刊论文

  • 发表日期：

    2025年8月18日

查看全文
  • Rahul Bhandari,
  • Purushottam Sharma,
  • Sanjay Singla,
  • Sandeep Kang

摘要

Agriculture is one of India's important sectors, impacting GDP and people's lives. This paper focuses on physical, chemical, and biological factors that affect soil fertility since it is the foundation of agricultural yield. However, conventional soil testing techniques are tedious and do not offer any information that can be used in precision farming. To this end, this study aims to use logistic regression, support vector machines, decision trees, random forests, and k-nearest neighbors to predict the soil fertility levels from the critical soil parameters. The key advance is to use a wide set of soil parameters, including macro and micronutrients, and soil physico-chemical properties like pH, OC, and EC to build predictive models. The results show that using the random forest algorithm is more efficient than other models, with an accuracy of 99%, followed by decision trees at 98%. These results show that machine learning can accurately predict soil fertility while minimizing costs. The study mentioned above aims to help farmers make informed decisions and enhance crop production by establishing the most effective predictive model for identifying potential soil conditions. The authors provide a set of recommendations to incorporate the findings and implications of the study into machine learning applications for improving agricultural yields and efficiency.