Research on Agricultural Product Unsold Risk Warning Method Based on Multi-Dimensional Data System and Combination Weighting Model

DOI: https://doi.org/10.62517/jbdc.202601332

Author(s)

Yuan Ren1, Hengxu Guo1, Yixuan Wang1, Yixin Yan1, Mingxu Li2

Affiliation(s)

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

Abstract

To address the problems of difficult multi-source data integration and strong subjectivity in risk quantification in agricultural product unsold warning, this paper constructs a multi-dimensional data system covering 14 data tables and 163,279 records, and proposes a seven-dimensional risk scoring model based on AHP-entropy weight combination weighting. Heterogeneous data across price, weather, yield, cost, circulation, and public opinion dimensions are collected from six types of data sources, with five preprocessing strategies improving data completeness from 87.3% to 99.6%, and a star schema is designed to organize them into a unified analytical framework. The analytic hierarchy process and the entropy weight method are linearly combined to determine weights, constructing a seven-dimensional risk scoring model covering price, supply, weather, cost, circulation, demand, and public opinion. Random forest 5-fold cross-validation achieves an AUC of 0.885 and a recall of 79.8%, and feature importance analysis shows that price risk contributes 44.88%, being the most critical predictor. Empirical analysis reveals that pork (CV=111.5%) and garlic (CV=30.1%) exhibit the most severe price volatility, October is the peak month for unsold events, and cold storage capacity shows only a weak positive correlation with unsold events (r=0.303), necessitating comprehensive measures combining production-marketing docking.

Keywords

Agricultural Product Unsold; Multi-Dimensional Data System; Star Schema; AHP-Entropy Weight Method; Combination Weighting; Risk Scoring; Random Forest CLC Number: TP311; F323 Document Code: A

References

[1] Ma X C, Song L L. Research on the causes and countermeasures of frequent unsold fresh agricultural products in China—A case study of vegetable and milk unsold. Contemporary Economic Management, 2015, 37(9): 59-62. [2] Ding H, Ren Y. How does rural e-commerce public service system construction solve the production-marketing dilemma of agricultural and sideline products—Practical evidence from the full-chain service system in Guizhou Province. Journal of Guizhou University of Finance and Economics, 2022(1): 66-74. [3] Yang R, Yang R C, Teng Q A. Research on risk grade evaluation of agricultural product unsold. Mathematics in Practice and Theory, 2020, 50(2): 318-328. [4] Deng Q, Wu Z C, Zhong Y. Construction of indicator system and risk grade assessment of agricultural product unsold under the background of big data—Research based on DEMATEL and grey prediction model. Journal of Zhengzhou University of Aeronautics, 2019(4): 11-25. [5] Liu J, Li Y F, Lei Y M, et al. Decision evaluation of agricultural product unsold risk grade based on triangular fuzzy analytic hierarchy process. Mathematics in Practice and Theory, 2019, 49(15): 299-305. [6] Xu S, Cai H M, Zhao L C, et al. A Bayesian decision tree algorithm model for risk prediction and assessment of agricultural product supply chain. Journal of Southwest University (Natural Science Edition), 2024, 46(3): 142-150. [7] Xu S W, Li Q C, Luan R P, et al. Deep learning intelligent prediction method for agricultural product market monitoring and early warning. Smart Agriculture, 2025, 7(1): 1-15. [8] Zhou H N, Liu P Z, Zhang H Q, et al. Research on garlic price prediction and early warning based on VMD-TCAN model. Transactions of the Chinese Society of Agricultural Engineering, 2026, 42(7): 239-248. [9] Wang R Z, Zhang X S, Wang M H. Agricultural product price prediction based on signal decomposition and deep learning. Transactions of the Chinese Society of Agricultural Engineering, 2022, 38(24): 256-265. [10] Zhuang J Y, Xu S W, Li Y, et al. Deep learning-based supply and demand prediction model for multiple agricultural products. Smart Agriculture, 2022, 4(2): 174-182. [11] Wang X, Zhu J Y, Qiu J X, et al. Livestock product risk analysis based on AHP-entropy weight method and Moran index. Transactions of the Chinese Society for Agricultural Machinery, 2026, 57(3): 337-348. [12] Hua X F, Liu H Q. Research on selection of "first mile" pre-cooling mode for agricultural products based on AHP-entropy weight method. Modern Management, 2025, 15(1): 165-179. [13] Li B W, Shao S H. Analysis of price volatility characteristics of agricultural products under different policy environments—Taking cotton and soybean as examples. World Agriculture, 2018(11): 100-107. [14] Liu J H, Ju R H. Research on the influence and stability of multi-period price weighting on agricultural product market supply decision—Based on second-order price-weighted cobweb model. Price: Theory & Practice, 2024(9): 133-144. [15] Ling L W, Xu M Q, Zhang X J. Analysis of influencing factors and prediction modeling of pork price volatility based on multi-dimensional association rules. Guangdong Agricultural Sciences, 2022, 49(12): 167-175. [16] Wang Z, Zhang W H, Fu J P, et al. Research on grape price prediction in Henan Province based on particle swarm optimization random forest algorithm. Modern Computer, 2024(7): 86-89. [17] He W J, Liao J P, Fu L L. Research on live pig price prediction and regulation mechanism based on random forest model. Prices Monthly, 2023(1): 21-27. [18] Guo W, Wu H R, Zhu H J, et al. Agricultural Big Data Governance: Key Technologies, Applications Analysis and Future Directions. Smart Agriculture, 2025, 7(3): 17-34. [19] Wang R Y, Strong D R. Beyond Accuracy: What Data Quality Means to Data Consumers. Journal of Management Information Systems, 1996, 12(4): 5-33. [20] Kimball R. The Data Warehouse Toolkit: Practical Techniques for Building Dimensional Data Warehouses. New York: John Wiley & Sons, 1996. [21] Zhou Z Y, Luo X W. Preliminary study on dimensional modeling of crop production management information data warehouse. Transactions of the Chinese Society of Agricultural Engineering, 2005, 21(11): 112-115. [22] Saaty T L. The Analytic Hierarchy Process: Planning, Priority Setting, Resource Allocation. New York: McGraw-Hill, 1980. [23] Shannon C E. A Mathematical Theory of Communication. Bell System Technical Journal, 1948, 27(3): 379-423. [24] Chen W, Xia J H. Multi-attribute decision making method integrating subjective and objective weight information. Systems Engineering—Theory & Practice, 2019, 39(6): 175-182. [25] Ezekiel M. The Cobweb Theorem. The Quarterly Journal of Economics, 1938, 52(2): 255-280. [26] Li H Y. Construction of fresh fruit and vegetable logistics system based on divergent cobweb model. Logistics Science and Technology, 2022, 45(3): 12-15. [27] Breiman L. Random Forests. Machine Learning, 2001, 45(1): 5-32