Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

AgroXAI: Explainable AI-Driven Crop Recommendation System for Agriculture 4.0

Ozlem Turgut, Ibrahim Kok, Suat Ozdemir

TL;DR

AgroXAI addresses the need for region-specific crop diversification in Agriculture 4.0 by integrating IoT-enabled edge computing with multiple ML classifiers and a suite of explainable AI methods. The system architecture places computation at the edge (physical/edge/fog/cloud layers) and uses RF, DT, KNN, SVM, LGBM, and MLP to classify suitable crops based on soil and weather features, while providing both global and local explanations via ELI5, SHAP, and LIME, plus counterfactuals for alternative crop recommendations. Results on a Kaggle agricultural dataset (2200 rows, 7 features, 22 crops) show RF achieving up to 99.24% accuracy, with SHAP/ELI5/LIME offering interpretable insights into feature importance such as Humidity and Rainfall, and counterfactuals enabling regionally feasible crop diversification. The work emphasizes explainability and regional trust, highlighting potential for sustainable, data-driven, region-aware agricultural decision-making and outlining considerations for security, privacy, and economic feasibility.

Abstract

Today, crop diversification in agriculture is a critical issue to meet the increasing demand for food and improve food safety and quality. This issue is considered to be the most important challenge for the next generation of agriculture due to the diminishing natural resources, the limited arable land, and unpredictable climatic conditions caused by climate change. In this paper, we employ emerging technologies such as the Internet of Things (IoT), machine learning (ML), and explainable artificial intelligence (XAI) to improve operational efficiency and productivity in the agricultural sector. Specifically, we propose an edge computing-based explainable crop recommendation system, AgroXAI, which suggests suitable crops for a region based on weather and soil conditions. In this system, we provide local and global explanations of ML model decisions with methods such as ELI5, LIME, SHAP, which we integrate into ML models. More importantly, we provide regional alternative crop recommendations with the counterfactual explainability method. In this way, we envision that our proposed AgroXAI system will be a platform that provides regional crop diversity in the next generation agriculture.

AgroXAI: Explainable AI-Driven Crop Recommendation System for Agriculture 4.0

TL;DR

AgroXAI addresses the need for region-specific crop diversification in Agriculture 4.0 by integrating IoT-enabled edge computing with multiple ML classifiers and a suite of explainable AI methods. The system architecture places computation at the edge (physical/edge/fog/cloud layers) and uses RF, DT, KNN, SVM, LGBM, and MLP to classify suitable crops based on soil and weather features, while providing both global and local explanations via ELI5, SHAP, and LIME, plus counterfactuals for alternative crop recommendations. Results on a Kaggle agricultural dataset (2200 rows, 7 features, 22 crops) show RF achieving up to 99.24% accuracy, with SHAP/ELI5/LIME offering interpretable insights into feature importance such as Humidity and Rainfall, and counterfactuals enabling regionally feasible crop diversification. The work emphasizes explainability and regional trust, highlighting potential for sustainable, data-driven, region-aware agricultural decision-making and outlining considerations for security, privacy, and economic feasibility.

Abstract

Today, crop diversification in agriculture is a critical issue to meet the increasing demand for food and improve food safety and quality. This issue is considered to be the most important challenge for the next generation of agriculture due to the diminishing natural resources, the limited arable land, and unpredictable climatic conditions caused by climate change. In this paper, we employ emerging technologies such as the Internet of Things (IoT), machine learning (ML), and explainable artificial intelligence (XAI) to improve operational efficiency and productivity in the agricultural sector. Specifically, we propose an edge computing-based explainable crop recommendation system, AgroXAI, which suggests suitable crops for a region based on weather and soil conditions. In this system, we provide local and global explanations of ML model decisions with methods such as ELI5, LIME, SHAP, which we integrate into ML models. More importantly, we provide regional alternative crop recommendations with the counterfactual explainability method. In this way, we envision that our proposed AgroXAI system will be a platform that provides regional crop diversity in the next generation agriculture.

Paper Structure

This paper contains 21 sections, 9 figures, 5 tables.

Table of Contents

  1. Introduction
  2. An Edge Computing-Based Explainable Crop Recommendation System (AgroXAI)
  3. System Model
  4. ML Models
  5. KNN
  6. RF
  7. DT
  8. SVM
  9. LGBM
  10. MLP
  11. Explainable AI Methods
  12. Explain Like I'm 5 (ELI5)
  13. SHapley Additive exPlanations (SHAP)
  14. Local Interpretable Model-agnostic Explanations (LIME)
  15. Counterfactual
  16. ...and 6 more sections

Figures (9)

  • Figure 1: Proposed Edge Computing-Based Explainable Crop Recommendation System (AgroXAI)
  • Figure 2: ELI5 global explanations for RF (Left) and LGBM (Right)
  • Figure 3: ELI5 local explanations for RF (Randomly selected sample test data: Nitrogen = 44, Phosphorus = 60, Potassium = 55, Temperature = 34.28046, Humidity = 90.555618, pH = 6.825371, Rainfall = 98.540474)
  • Figure 4: ELI5 local explanations for LGBM (Randomly selected sample test data: Nitrogen = 44, Phosphorus = 60, Potassium = 55, Temperature = 34.28046, Humidity = 90.555618, pH = 6.825371, Rainfall = 98.540474)
  • Figure 5: SHAP global explanations for RF (Left) and LGBM (Right)
  • ...and 4 more figures