Added Crop Rotation Recommendation system

[J-B-Mugundh]

Description

Added the Crop Rotation Recommendation system, which suggests the optimal next crop based on previous crop, soil type, moisture level, and nutrient content (Nitrogen, Phosphorus, Potassium). This feature uses a machine learning model to assist in sustainable farming practices by predicting the best crop rotation.

Issue Resolved

Resolves #106

Changes Made

• Integrated the crop recommendation model.
• Added a new section with input fields for Previous Crop, Soil Type, Moisture Level, Nitrogen (N), Phosphorus (P), and Potassium (K).
• Implemented mappings for crop and soil types.
• Added prediction logic that returns the recommended crop.
• Added model's documentation and explanation for the new feature.

Screenshots or Videos

Additional Details

The dataset is not already existing one in the internet! Rather, it was synthesized specifically for this purpose from various internet sources.

Checklist

• [x] My code follows the current project structure
• [x] I have thoroughly reviewed and updated the requirements.txt file to include any new packages
• [x] The predict.py file includes a properly implemented model_details() function, or the notebook contains this function to print a detailed model report. The model will not be accepted without this function, as it is essential for generating the necessary model details.
• [x] I have added relevant tests (if necessary).
• [x] I have added comments in the code where needed.
• [x] This PR is submitted under Hacktoberfest.
• [x] This PR is submitted under GirlScript Summer of Code (GSSoC-Extd).

[yashasvini121]

I'm closing this PR for now. Please improve the model and submit a new PR once ready.

[J-B-Mugundh]

@yashasvini121 Actually, the dataset was synthesized.. I think that's why this issue is.. could u suggest any idea on how to improve this?

[yashasvini121]

First, try to find a relevant dataset. If that's not feasible, dedicate time to researching the topic. Your dataset was highly unbalanced, with some crops appearing 10K times while others only in the thousands. Once the synthetic dataset is generated, analyze it using visualizations and other methods to ensure it accurately reflects real-world conditions.