AI & softwareResearch findings
SkinAdvisor · Development of an AI-Based Decision Framework for Skin Sales Forecasting
Original published 2026.11.11 13:36 KSTAdded to Blackdeck 2026.11.15 21:29 KST
What has changed?
A multi-modal Transformer model integrates player preference and skin relationship data to forecast sales trajectories for unreleased skins.
Key points
- Introduces an AI-driven framework to systematize skin production planning based on sales forecasts.
- Leverages multi-modal Transformer to incorporate player preferences and skin relationships in forecasting.
- Achieves ~20% RMSE reduction over baseline models in offline experiments.
Connect it to your work
Forecast accuracy improves by ~20% over baseline models; seasonal splitting prevents temporal leakage in embeddings.
What remains unproven
This research note is based on an abstract; check the original for test conditions and application scope.
Originals used for this note
Links open the original publisher in a new tab.