Inspiration: This project was inspired by the growing need to solve problems in the agricultural supply chain, such as fluctuating crop prices, high levels of waste, and difficulties farmers face in finding proper storage. By using AI and machine learning, we aim to help farmers, warehouse owners, and flower sellers get real-time insights to manage crop storage, reduce waste, and predict future price trends, ultimately making agriculture more efficient and profitable.
What it does: Our website integrates a SARIMA-NARNET hybrid model to provide accurate predictions of crop prices and production. It allows farmers to list their crops for storage, and warehouse owners can easily search and store them. The platform also supports price negotiations, tracks shipments, and reduces agricultural waste by converting it into organic manure. Additionally, it connects surplus food from urban areas to regions in need, creating an efficient food distribution network.
How we built it: We built the SARIMA-NARNET hybrid model using Python, combining the time-series forecasting ability of SARIMA with the nonlinear prediction strengths of NARNET. The website was developed using React.js for the frontend and Node.js for the backend, while MySQL was used for managing the database. We also used TensorFlow for the neural network implementation. The model was trained using agricultural datasets to improve the accuracy of the predictions, and the website was designed to provide an easy-to-use experience for managing crop storage and tracking.
Challenges we ran into: Some major challenges we faced were in fine-tuning the SARIMA-NARNET model to get accurate predictions, especially with nonlinear trends in agricultural data. Another challenge was integrating the model with the website to ensure that real-time price updates were provided smoothly. Developing a user-friendly interface while managing complex backend functionalities and creating an efficient food logistics system were also challenging aspects of the project.
Accomplishments that we're proud of: We are proud to have successfully developed a prediction model with 90% accuracy, which was a significant achievement considering the complexity of the data. Additionally, 65% of the website development is complete, including important features like crop listing, deal negotiation, and shipment tracking. Another highlight was successfully integrating AI-driven waste reduction and surplus food distribution systems, which contribute to sustainability and help tackle food insecurity.
What we learned: Through this project, we learned how to deal with the complexities of agricultural forecasting, combining time-series analysis with neural networks. We also gained experience in addressing scalability issues when integrating machine learning models into a website. Additionally, we learned about logistics optimization and how AI can be applied to improve agriculture and reduce waste.
What's next for Demand Forecasting in the Field of Agriculture: Our next steps include finishing the remaining 35% of the website and refining the user interface to make it more intuitive. We also plan to improve the accuracy of our prediction model further. In the future, we aim to add advanced AI tools for monitoring crop health and predicting pest infestations. Additionally, we will expand our surplus food distribution network by collaborating with more food organizations and NGOs to have a larger impact on food security.
Log in or sign up for Devpost to join the conversation.