Abstract:
The project aims to develop a user-friendly web application to predict suitable crops
based on weather conditions in Pakistan's agriculture sector. The application will also
provide recommendations on alternative crops that can survive adverse climatic
conditions in addition to the forecast of primary crop selection.
In addition, it will provide farmers with information on the required soil types and
optimal levels of fertilizers for each recommended crop, to provide them with tailored
guidance. Our study enhances the methods of crop prediction, enables farmers to
benefit from valuable decision-making tools, and increases agricultural productivity
by using a variety of machine-learning algorithms including Logistic Regression,
Random Forest, K-Nearest Neighbours (KNN), Naive Bayes, Decision Tree, and
Support Vector Machine (SVM)
Following Agile methodology, we iteratively refine software construction, leveraging
collaboration and adaptability. We also apply the Agile principles to enhance software
functionality through the integration of new data in machine learning. Given factors
like temperature, humidity, and rainfall, our methodology involves full data collection,
thorough cleaning, and pre-processing. In the end, we are aiming to create a simple
interface for widespread access to information and decision-making in Pakistan's
agricultural landscape so that it can contribute to the financial stability of the country.
Our Project predicts the crop yielding results with approx. 90% accuracy. It also
provides alternative ways to the farmers for better results in future.
