Soiligator 💦⏳ Soil Analysis and Irrigation Prediction

Overview

Soiligator is an advanced machine learning project designed to optimize irrigation management by predicting whether irrigation is necessary based on environmental and soil-related data. Leveraging feature engineering and robust predictive models, Soiligator provides actionable insights that improve agricultural efficiency and sustainability.

Key Features

• Predictive Models: Utilizes Logistic Regression, Random Forest, and Support Vector Machine (SVM) algorithms for accurate irrigation predictions.
• Feature Engineering: Incorporates non-linear interaction terms and outlier handling for enhanced model performance.
• Scalable Design: Easily extendable to include additional features like soil type and crop variety.
• Data Resilience: Designed to handle label noise and outliers, ensuring robustness in real-world applications.

Table of Contents

1. Overview
2. Key Features
3. Installation
4. Usage
5. Data Description
6. Model Training and Evaluation
7. Results
8. Future Work

Installation

To use this project, install the required Python packages with the following command:

pip install -r requirements.txt

Key Dependencies:

• pandas: Data manipulation and analysis
• numpy: Numerical operations
• matplotlib & seaborn: Data visualization
• scikit-learn: Machine learning model training and evaluation

Alternatively, install the libraries manually:

pip install pandas numpy matplotlib seaborn scikit-learn

Usage

Data Loading

Start by loading the dataset modified_irrigation_dataset.csv, which includes:

• Moisture: Soil moisture content.
• Temperature: Ambient temperature.
• Humidity: Air humidity level.
• Irrigation_Needed: Target label indicating whether irrigation is required.

Running the Code

The implementation is available in a Jupyter Notebook: soil_analysis.ipynb. Execute the cells sequentially to:

1. Load and preprocess the dataset.
2. Engineer additional features.
3. Train machine learning models.
4. Evaluate and compare model performance.

Data Description

The dataset comprises features representing soil and environmental conditions:

• Moisture: Measures the water content in the soil (0–100%).
• Temperature: Ambient temperature in degrees Celsius.
• Humidity: Air humidity as a percentage (0–100%).

Engineered Features:

• Moisture_Temp_Interaction: Interaction term between soil moisture and temperature to capture non-linear effects.
• Humidity_Squared: Non-linear transformation of humidity to account for atmospheric retention properties.

Data Challenges:

• Outliers: Synthetic outliers introduced in 5% of the data to test model resilience.
• Label Noise: Added noise to 5% of target labels to simulate real-world conditions.

Model Training and Evaluation

Preprocessing

• Outlier Handling: Removes or neutralizes extreme values.
• Feature Scaling: Standardizes features using StandardScaler for optimal model performance.
• Train-Test Split: Splits the data into 80% training and 20% testing subsets.

Models Used:

1. Logistic Regression: A baseline model for binary classification.
2. Random Forest Classifier: An ensemble learning model for handling complex patterns.
3. Support Vector Machine (SVM): A robust classifier for high-dimensional data.

Evaluation Metrics:

• Accuracy: Overall correctness of predictions.
• Confusion Matrix: Breakdown of true positives, false positives, true negatives, and false negatives.
• ROC Curve and AUC Score: Measures the model's ability to distinguish between classes.
• Precision-Recall Curve: Highlights performance in handling imbalanced data.
• Classification Report: Includes precision, recall, F1-score, and support.

Results

Model Comparison:

• Logistic Regression: Achieved baseline performance with moderate accuracy.
• Random Forest: Outperformed other models, achieving high accuracy and robustness to noise and outliers.
• SVM: Demonstrated strong performance on standardized features but required longer training times.

Visualization:

• Confusion Matrix: Provided for each model to analyze prediction errors.
• ROC Curves: Highlighted the trade-offs between sensitivity and specificity.
• Precision-Recall Curves: Demonstrated model effectiveness on imbalanced datasets.

Future Work

1. Hyperparameter Tuning: Optimize models using Grid Search or Random Search to improve accuracy.
2. Feature Expansion: Include additional predictors such as:
   • Soil type
   • Crop type
   • Real-time weather forecasts
3. Time-Series Analysis: Incorporate temporal data to predict irrigation needs over time.
4. Deployment: Package the model into a web or mobile application for practical use by farmers and agricultural experts.

---

Contribution

Contributions are welcome! Please fork the repository, make your changes, and submit a pull request. For any queries, feel free to contact the project owner.