감자_20260828_message_mp4_narrow.mp4
2026-08-28 · 21.7 MB
DUNG
VERSION 1.0Predicting the future of agricultural prices
Data source: NongNet https://www.nongnet.or.kr/DUNG forecasts price movements of agricultural products traded in Korea's wholesale markets to support decisions by producers, buyers, and institutions. It provides past prices alongside future outlooks so they can be used in real decisions.
Different models are applied depending on the goals of producers, buyers, and institutions. Short, medium, and long forecast horizons support everything from near-term shipping to mid/long-term planning. Based on recent transactions, DUNG recommends shipping and purchase timing and estimates the additional profit or loss from following that recommendation.
Forecasts are provided by growing region, market, product, variety, and grade. Agricultural processors, farmers, producer groups, local agricultural cooperatives, APCs, wholesale associations, and food companies can use it for shipping, storage, volume, and sales planning. Wholesale corporations and food processors (buyers) get purchase timing and market-level distribution volume information.
DUNG combines DNN/Attention/Cascade-based neural networks, GRU (an RNN family model), and multiple tabular data models through an orchestrator. Combining the judgments of different models is designed to minimize the bias of any single model, and model details and performance metrics are published openly.
Starting with 20 products, coverage will expand to 100+ by the end of 2026. The service is free of personal data collection and ads, and raw forecast data is available upon request.
Original forecast data and analysis files, free to download for anyone — no personal information, no ads.