What does this research mean for the field?

Historical sales data, particularly one-day sales lags, provide a stronger predictive signal for e-commerce retail sales forecasting than external weather variables, with Stacked LSTM models achieving the highest forecasting accuracy. Novelty: ClaimNovelty.INCREMENTAL. Consensus alignment: ConsensusAlignment.NEUTRAL.

What question did this study set out to answer?

To evaluate an interpretable machine learning framework for forecasting daily retail sales revenue and assess the added predictive value of external weather features compared to historical sales data.

September 15, 2026LogisticsOpen Access

A Machine Learning Framework: E-Commerce Retail Sales Forecasting with External Weather Features

View Full Paper

Ask AI

Bookmark

Authors

MSMuhammad Zaka ShaheryarUniversity of Toronto MAMonzer AlharairiUniversity of Toronto SASaman Hassanzadeh AminUniversity of Toronto

Discussion

Member takes

Implication

Predictive modeling evaluation finds stacked LSTM optimizes daily revenue forecasting in e-commerce, indicating historical sales outweigh external weather factors.

Key Points

To evaluate an interpretable machine learning framework for forecasting daily retail sales revenue and assess the added predictive value of external weather features compared to historical sales data.
Integrated daily location-level sales data with weather observations and temporal lag variables.
Trained and evaluated Random Forest, Extreme Gradient Boosting, CatBoost, and Stacked Long Short-Term Memory (LSTM) architectures using MAE, RMSE, sMAPE, and R2 metrics.
Stacked LSTM achieved the strongest forecasting accuracy on test data (MAE = 28.50, RMSE = 57.04, sMAPE = 5.82%, R2 = 0.972).
Feature-importance analysis demonstrated that historical sales metrics, especially a one-day sales lag, provided a substantially stronger predictive signal than individual weather variables.

Cite This Study

Shaheryar et al. (2026) studied this question.

synapsesocial.com/papers/6aa913ba9013453be30a1cb4 https://doi.org/10.3390/logistics10090214

View Full Paper

Ask AI

Bookmark

Implication

Predictive modeling evaluation finds stacked LSTM optimizes daily revenue forecasting in e-commerce, indicating historical sales outweigh external weather factors.

A Machine Learning Framework: E-Commerce Retail Sales Forecasting with External Weather Features

Discussion

Member takes

Implication

Key Points

Cite This Study

A Machine Learning Framework: E-Commerce Retail Sales Forecasting with External Weather Features

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider