Scholarlink

Currently, predicting the prices of agri-horticultural commodities, such as pulses and vegetables like onions and potatoes, is challenging. Stakeholders in this project include farmers, traders, and policymakers who can utilize accurate commodity price predictions for improved decision-making. The Department of Consumer Affairs, India, generates data from more than 550 reporting centers to track price movements of many food stuffs. They also maintain buffer stocks of some pulses and onions. Based on historical data, seasonality, and market intelligence, ARIMA model is being used to give immediate insights. But, ARIMA model is a statistical approach and uses only linear data. There will be a lot of nonlinearity in the real-time data. Long Short-Term Memory (LSTM) is being used to handle this issue. It uses complex relationship between the data to give predictions. In this idea, we have considered potato, onion, tur, masur, wheat, urad wholesale price index dataset from www.data.gov.in and other sources. The information about weather is also considered for precise prediction. This data is collected using OpenWeatherMap API. The prediction score of most of the products is around 95.3%. This is a way better than the existing approaches.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Integrating Real-Time Weather Data with LSTM for Accurate Price Forecasting of Agri-Horticultural Commodities

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • I. Kali Pradeep,
  • Bhagyasri Appana,
  • Boddu Amrutha,
  • Bolla Geethika Sridevi,
  • Gundavarapu Naveen

摘要

Currently, predicting the prices of agri-horticultural commodities, such as pulses and vegetables like onions and potatoes, is challenging. Stakeholders in this project include farmers, traders, and policymakers who can utilize accurate commodity price predictions for improved decision-making. The Department of Consumer Affairs, India, generates data from more than 550 reporting centers to track price movements of many food stuffs. They also maintain buffer stocks of some pulses and onions. Based on historical data, seasonality, and market intelligence, ARIMA model is being used to give immediate insights. But, ARIMA model is a statistical approach and uses only linear data. There will be a lot of nonlinearity in the real-time data. Long Short-Term Memory (LSTM) is being used to handle this issue. It uses complex relationship between the data to give predictions. In this idea, we have considered potato, onion, tur, masur, wheat, urad wholesale price index dataset from www.data.gov.in and other sources. The information about weather is also considered for precise prediction. This data is collected using OpenWeatherMap API. The prediction score of most of the products is around 95.3%. This is a way better than the existing approaches.