Research on E-Commerce Inventory Sales Forecasting Model Based on ARIMA and LSTM Algorithm DOI Creative Commons
Chenyang Wang, Junsheng Wang

Mathematics, Год журнала: 2025, Номер 13(11), С. 1838 - 1838

Опубликована: Май 31, 2025

Accurate forecasting is critical for effective warehouse network planning and inventory management in e-commerce. This study tackles these challenges by applying a differentiated strategy over three-month period. The Autoregressive Integrated Moving Average (ARIMA) model used monthly predictions, while the Long Short-Term Memory (LSTM) neural employed daily sales forecasts. Experimental validation across 350 product categories demonstrates efficacy of this approach. ARIMA effectively captured dynamic trends (e.g., Category 1 showing gradual increases; 91 depleting from 3824 to 0). Concurrently, LSTM successfully modeled complex fluctuations 61 peaking at 3693 units on 21 July; 31 consistently recording zero sales). dual-model strategy, leveraging complementary strengths relatively stable series volatile patterns, provides robust, data-driven basis optimizing resource category allocation. Furthermore, visualization categorized forecast results reveals distinct distribution thereby enabling enterprises refine strategies with greater precision, leading reduced redundant space investment improved allocation efficiency. Future research will focus incorporating multivariate interactions further enhance practicality predictive power.

Язык: Английский

Research on E-Commerce Inventory Sales Forecasting Model Based on ARIMA and LSTM Algorithm DOI Creative Commons
Chenyang Wang, Junsheng Wang

Mathematics, Год журнала: 2025, Номер 13(11), С. 1838 - 1838

Опубликована: Май 31, 2025

Accurate forecasting is critical for effective warehouse network planning and inventory management in e-commerce. This study tackles these challenges by applying a differentiated strategy over three-month period. The Autoregressive Integrated Moving Average (ARIMA) model used monthly predictions, while the Long Short-Term Memory (LSTM) neural employed daily sales forecasts. Experimental validation across 350 product categories demonstrates efficacy of this approach. ARIMA effectively captured dynamic trends (e.g., Category 1 showing gradual increases; 91 depleting from 3824 to 0). Concurrently, LSTM successfully modeled complex fluctuations 61 peaking at 3693 units on 21 July; 31 consistently recording zero sales). dual-model strategy, leveraging complementary strengths relatively stable series volatile patterns, provides robust, data-driven basis optimizing resource category allocation. Furthermore, visualization categorized forecast results reveals distinct distribution thereby enabling enterprises refine strategies with greater precision, leading reduced redundant space investment improved allocation efficiency. Future research will focus incorporating multivariate interactions further enhance practicality predictive power.

Язык: Английский

Процитировано

0