It a web based application for crop prediction and suggesting fertilizer for a crop based on Aritifical Intelligence algorithm called Constraint Satisfaction Problem (CSP)
Description
Website Walkthrough
Technologies Used
CSP
Intelligent Backtracking
Naive Bayes
Future works
Given many parameters and different constraints, the system must recommend the suitable crop to be grown such that it should give maximum profit to the farmers as well as meet the market demand. The system also has a feature to suggest a fertilizer based on the nutrient contents present in a particular soil and type of crop to be grown in that particular soil.
- HTML - Interactive UI
- CSS - Styling
- Python - To implement CSP alogorithm
- Flask - Web Framework
- Matplotlib - Data Visualization
- Sklearn - Data Modelling
It consist of 3 elements:
- Domains
- Variables
- Constraints
In which :
- Variables: A list containt crops,
- Domains: A set of 10 crops,
- Constraints: Temperature,Precipitation,Humidity etc..,
- Dataset: It has data taken from government and synthetic data so it has an accuracy of 89% by providing proper dataset it can able to provide crops with an accuracy of 99.5%
It can be implemented in 3 ways:
- Forward Chaining
- Backtracking
- Intelligent Backtracking
In which we have used intelligent Backtracking which is discussed later
It get the constraints from the user and based on it finds the specific crops that satisfy that particular constraint and it will be added in the variables once if a contraint is not satisified it add the previous crop in the variables and does the intelligent backtracking and begins with new crop.
We have used Naive bayes model to detect the accuracy of the data collected from various websites to make sure that, the prediction made with that particular dataset is mostly accurate.
To include a feature to detect disease by uploading an image of the infected plant .