Education, Science, Technology, Innovation and Life
Open Access
Sign In

Pricing and Replenishment Decision of Vegetable Goods Based on LSTM and XG-Boost Models

Download as PDF

DOI: 10.23977/infse.2023.041012 | Downloads: 15 | Views: 583

Author(s)

Zhengyi Luo 1, Xin Zou 1

Affiliation(s)

1 School of Electronics and Information, Nanchang Institute of Technology, Nanchang, China

Corresponding Author

Zhengyi Luo

ABSTRACT

In general, vegetable commodities do not have a long shelf life, and along with the increase in time the quality of vegetable commodities will be reduced, so many super if the day is not sold out, basically cannot be sold again such commodities, which invariably increases the rate of loss. To this end, this paper analyzes the relevant sales data and develops appropriate pricing and replenishment decisions. First of all, the distribution analysis of vegetables in various categories of goods, based on the Kendall correlation coefficient test for the consistency of the single product test, and then the correlation between the two categories of correlation analysis. Then, a mathematical model is constructed to maximize the revenue of the superstore, using the LSTM time-series prediction model to predict the wholesale price of each category of vegetables in the coming week based on the historical wholesale price, the GBDT sales volume prediction model based on the unit price and wholesale price, the objective function of maximizing the revenue of the superstore is set up, and the total revenue of each category of vegetables in the coming week is obtained through Monte Carlo algorithm solving. Finally, in order to meet the requirements of the total number of individual items and the minimum display quantity, the LightGBM time series prediction model is constructed on the historical wholesale price data to predict the wholesale price of the vegetable category on a single day and establish the objective model for maximizing the revenue of the superstore, which is determined by the daily sales quantity, the sales unit price, and the difference of the wholesale price together with the wastage rate. The proposed model has high solution efficiency and optimality.

KEYWORDS

Kendall Correlation Coefficient, LSTM, XG-Boost, Goal Planning Model, Monte Carlo Algorithm

CITE THIS PAPER

Zhengyi Luo, Xin Zou, Pricing and Replenishment Decision of Vegetable Goods Based on LSTM and XG-Boost Models. Information Systems and Economics (2023) Vol. 4: 83-92. DOI: http://dx.doi.org/10.23977/infse.2023.041012.

REFERENCES

[1] Hou X, Fang G, Improved ID3 algorithm based on Kendall coefficient 1, Science and Technology Innovation, 2021.
[2] Li S, Research on accurate prediction of heavy overload of transformer based on GBDT algorithm, Modern Information Technology, 2023(4).
[3] Zha W, Dong Y, Jiang Z, Liu Y, A prediction model of manganese content at the endpoint of vacuum self-consumption ingot casting based on GBDT algorithm, Metallurgy in China, 2023 (7).
[4] Chen X, Guo C, Yuan C, Huang W, Liu Z, Intelligent Air Quality Monitoring and Governance Department Based on Monte Carlo Algorithm, Information Recorded Materials, 2022(9).
[5] Pan X, Xie Z, Wang S, Optimal decision making for fresh food superstore preservation efforts and pricing considering loss aversion, Highway Transportation Science and Technology, 2022(6).
[6] Gao Y, Wang W, Wang J. An integrated fuzzy hierarchical classification of LightGBM food safety risk early warning model: a case study of meat products, Food Science, 2021(1).
[7] Zhang X, He L, Zheng J. Research on Sales Forecasting Based on Gray Correlation Analysis and XGBoost Model. Software Journal, 2020, 19(09).

Downloads: 12853
Visits: 257807

All published work is licensed under a Creative Commons Attribution 4.0 International License.

Copyright © 2016 - 2031 Clausius Scientific Press Inc. All Rights Reserved.