Overview / Usage

Inspiration:

The UN estimates that every year, almost half of the fruit and vegetables produced worldwide are lost or wasted, resulting in a substantial loss of resources and energy. This is why the motivation for a crop forecast machine learning model is to encourage farmers to produce more and to fight against human and animal starvation. This initiative strives to maximise crop output and eliminate waste in order to create a more sustainable and effective food system for everyone. It does this by utilising cutting-edge technology and data analysis. With the help of this model, farmers may decide on their crops with knowledge, resulting in higher yields and less need for excessive use of resources like water, fertiliser, and pesticides. Ultimately, this project has the potential to play a significant role in improving food security and reducing hunger on a global scale.

Methodology / Approach

The Crop Recommendation application is an advanced tool that assists farmers in making informed decisions about which crops to cultivate. The application utilizes machine learning algorithms to deliver personalized crop recommendations based on the soil data provided by the user. By entering information such as soil pH levels, nutrient content, and moisture levels, farmers can receive recommendations for crops that are most suitable for their soil conditions. The application takes into account various factors, including regional climate, weather patterns, and historical crop yields, to provide accurate and customized predictions. This enables farmers to focus on crops that are most likely to thrive on their land, thereby reducing wasted resources and improving yields. Overall, the Crop Recommendation application is a valuable tool for farmers seeking to optimize their crop production and increase their profits.

​

Based on the results of the model evaluation, the Crop Prediction project achieved impressive accuracy scores across different machine learning algorithms. The support vector machine (SVM) algorithm had the lowest accuracy score of 0.1068, while the logistic regression algorithm achieved a much higher accuracy score of 0.9523. The random forest (RF) algorithm achieved an even higher accuracy score of 0.9841, and the XGBoost algorithm had the highest accuracy score of 0.9932. These high accuracy scores demonstrate the effectiveness of the machine learning algorithms in providing personalized crop recommendations to farmers.

Technologies Used

Intel oneDAL

Intel devcloud

Machine Learning

Sklearn

Numpy

Pandas

Matplotlib

Seaborn

Repository

https://github.com/JoelJJoseph/CROP_PREDICTION