Fadhilahnur/Agriculture-Crop-Recommendation-System

By using external factors on fertilizers composition, climate change and geographical features, I used classification method in machine learning model by using Random Forest in Python to create this agriculture crop recommendation system as an innovative approach to plan and improve crop yield. Data is stored in MSSQL database, then connect Python to the database to code the machine learning model. The user interface (UI) is created by using Python TKinter. The calculation from Random Forest is pack using pickle module in Python, where it can be connected to UI to called the result for crop yield suggestion when user key in their data. This dataset contains 22 types of crops yield with details on Nitrogen, Potassium, Phosphorus, temperature, humidity, pH Level and rainfall. This model is tested using Decision Tree, Naïve Bayes, Support Vector Machine, Logistic Regression and Random Forest. Out of this 5 models, Random Forest has the highest accuracy. This approach may be able to reduce vulnerability in agriculture landscape and questioned raised on meeting global food demand sustainability. This also can be used as solution for young farmers to fully optimize their planning on what crop they can plant based on the current geographical features and mineral composition.