Multi-source Data-driven Cause Identification and Risk Early Warning for Unsaleable Agricultural Products

DOI: https://doi.org/10.62517/jike.202604404

Author(s)

Liu Zhanhong1, Chen Mengfei1, Wang Zhanyu1, Zhang Jiahang1, Li Mingxu2

Affiliation(s)

1School of Artificial Intelligence and Big Data, Henan University of Technology, Zhengzhou, Henan, China 2IFLYTEK Co., Ltd., Hefei, Anhui, China

Abstract

Farm products may not be sold due to a mismatch in harvesting time, buyer demand, transportation, information and market organization; excessive output is not the only reason for this problem. We take the monthly data of eleven products in nine Henan cities from 2019 to 2025 and use a city-product-month working unit. Analyze the construction of the seven-factor risk profile, derive a ten-variable index with entropy weighting, attribute observations to five relative risk groups, and examine five categories of causes. The Cost Structure is relatively rigid. Chinese cabbage has a high cost risk and resulted in a loss in 2023 and 2024; this combination also occurred for pork in 2023. Maize was profitable and low-risk in 2023, but it became loss-making and high-risk the following year. Diesel and road-freight indicators were relatively high in 2022, but there were still some fluctuations month-to-month. Online attention is also uneven, and chili has 73 anomalies and is the most numerous in the sample. City results do not have a single channel story. Shangqiu has high e-commerce penetration and a relatively high reported frequency; Sanmenxia also has high penetration but a much lower frequency. Only a few product-year cases of the cost signal have been identified: Chinese cabbage in 2023-2024, pork in 2023, and maize in 2024. Freight pressure was highest in 2022. The two cities do not have the same coverage; although both are highly online, the frequency of reported cases in Shangqiu and Sanmenxia is different. Therefore, these records were selected for inspection but not included in the closed-won. Search anomalies are used to identify which records need to be verified. They are not considered evidence of a transaction or that a product has gone out of stock. Therefore, the corresponding response will be taken to address the bottleneck in the data. A cost measure is suitable for one product, organisation of transportation needs to take place before a sensitive marketing window, and an online signal should be compared with market data prior to action. Frequent Combinations can help to find cases for investigation, but they do not establish causality or predict a particular person's behaviour.

Keywords

Unsaleable Agricultural Products; Causes of Unsaleable Surplus; Multi-Source Data; Risk Early Warning; E-Commerce-Enabled Agriculture; Henan Province.

References

[1] Hao L, Lee K S. The impact of economic policy uncertainty on agricultural prices: Evidence from China. Asia and the Global Economy, 2024, 4(2): 100100. DOI: 10.1016/j.aglobe.2024.100100. [2] Luo M, Zhou G, Xu H. A differential game model research on dynamic pricing and coordination of fresh agricultural products supply chain based on freshness. Economic Research-Ekonomska Istraživanja, 2023, 36(2): 2177696. DOI: 10.1080/1331677X.2023.2177696. [3] Zhu H, Liu C, Wu G, Gao Y. Cold Chain Logistics Network Design for Fresh Agricultural Products with Government Subsidy. Sustainability, 2023, 15(13): 10021. DOI: 10.3390/su151310021. [4] Guo J, Jin S, Zhao J, Li Y. E-commerce and supply chain resilience during COVID-19: Evidence from agricultural input e-stores in China. Journal of Agricultural Economics, 2023, 74(2): 369-393. DOI: 10.1111/1477-9552.12518. [5] Fu L, Zhang H. Analysis of factors influencing small-scale agricultural product prices from the perspective of the online public—a case study of China. Frontiers in Sustainable Food Systems, 2024, 8: 1355853. DOI: 10.3389/fsufs.2024.1355853. [6] Liu S, Jin Y, Zheng F. Did product certification and e-commerce benefit agricultural producers in China?. Food Policy, 2024, 127: 102688. DOI: 10.1016/j.foodpol.2024.102688. [7] Jiang L, Zhang Y. Modeling Urban Expansion and Agricultural Land Conversion in Henan Province, China: An Integration of Land Use and Socioeconomic Data. Sustainability, 2016, 8(9): 920. DOI: 10.3390/su8090920. [8] Tan Y, Zeng H. Price transmission, reserve regulation and price volatility. China Agricultural Economic Review, 2019, 11(2): 355-372. DOI: 10.1108/CAER-04-2017-0062. [9] Ling L, Zhang D, Chen S, Mugera A W. Can online search data improve the forecast accuracy of pork price in China?. Journal of Forecasting, 2020, 39(4): 671-686. DOI: 10.1002/for.2649. [10] Lu H, Sun G, Zhou L, Zhang W, Zou J. The impact and mechanism of E-commerce of agricultural products on nongrain production of cultivated land in rural China. Journal of Environmental Management, 2025, 373: 123639. DOI: 10.1016/j.jenvman.2024.123639. [11] Liu Z, Zhao K, Ma J, Wang C. A Price Index Model for Road Freight Transportation and Its Empirical analysis in China[C]. MATEC Web of Conferences, 2017, 100: 05017. DOI: 10.1051/matecconf/201710005017. [12] Tian X, Yi F, Yu X. Rising cost of labor and transformations in grain production in China. China Agricultural Economic Review, 2019, 12(1): 158-172. DOI: 10.1108/CAER-04-2018-0067. [13] Zhang T, Feng Q. Multidimensional Risk Assessment of China’s Grain Supply Chain with Entropy Weight TOPSIS Method. Procedia Computer Science, 2024, 242: 853-858. DOI: 10.1016/j.procs.2024.08.213. [14] Wang H, Lu L, Liu Z, Sun Y. Study on measurement and prediction of agricultural product supply chain resilience based on improved EW-TOPSIS and GM (1,1)-Markov models under public emergencies. PLOS ONE, 2025, 20(5): e0321248. DOI: 10.1371/journal.pone.0321248.