Attention based Multi-Modal New Product Sales Time-series Forecasting
Trend driven retail industries such as fashion, launch substantial new products every season. In such a scenario, an accurate demand forecast for these newly launched products is vital for efficient downstream supply chain planning like assortment planning and stock allocation. While classical time-series forecasting algorithms can be used for existing products to forecast the sales, new products do not have any historical time-series data to base the forecast on. In this paper, we propose and empirically evaluate several novel attention-based multi-modal encoder-decoder models to forecast the sales for a new product purely based on product images, any available product attributes and also external factors like holidays, events, weather, and discount. We experimentally validate our approaches on a large fashion dataset and report the improvements in achieved accuracy and enhanced model interpretability as compared to existing k-nearest neighbor based baseline approaches.
PDF AbstractDatasets
Results from the Paper
Ranked #1 on New Product Sales Forecasting on VISUELLE2.0 (using extra training data)
Task | Dataset | Model | Metric Name | Metric Value | Global Rank | Uses Extra Training Data |
Benchmark |
---|---|---|---|---|---|---|---|
New Product Sales Forecasting | VISUELLE | Explainable Cross-Attention Multimodal RNN | MAE | 32.1 | # 5 | ||
New Product Sales Forecasting | VISUELLE2.0 | Explainable Cross-Attention Multimodal RNN | MAE | 0.99 | # 1 | ||
Short-observation new product sales forecasting | VISUELLE2.0 | Explainable Cross-Attention Multimodal RNN | 10 steps MAE | 0.94 | # 1 | ||
1 step MAE | 0.96 | # 1 |