[KongchangAI]
Model销量预测模型

SkinCast

A multimodal Transformer sales forecasting model based on four-encoder fusion and global-local positional encoding. It can forecast day-by-day sales trajectories over 180 days for unreleased skins.

Timeline (last 90 days)

Nov 13

SkinCast is the core sales forecasting model in the SkinAdvisor framework, capable of predicting the day-by-day sales trajectory of unreleased skins over 180 days.

Unverified50%
Nov 13

SkinCast contains four encoders: a pretrained Informer that processes player preference time series; a pretrained Player-Hero-Skin graph embedder that captures relational priors; and two lightweight Transformers that encode skin attributes and release context, respectively.

Unverified50%
Nov 13

SkinCast's fused features pass through a shared backbone with local-plus-global day-x positional encoding, then through an N-layer Informer decoder that outputs the forecast sequence in a single non-autoregressive pass.

Unverified50%
Nov 13

Non-autoregressive decoding requires only one forward pass at inference time, significantly reducing latency, and is one of the key designs enabling SkinCast to achieve millisecond-level response on CPU.

Unverified50%
Nov 13

Offline experiments show that SkinCast reduces RMSE by about 20% compared with the strongest baseline, Informer, and by about 60% compared with classical methods.

Unverified50%
Nov 13

Stress tests conducted on a 16-core CPU server confirmed that SkinCast's p99 latency is below 8 milliseconds and its peak throughput exceeds 1100 rps.

Unverified50%
Nov 13

The paper points out that pretrained embeddings may leak temporal information, requiring a season-based split to prevent data leakage.

Unverified50%
Nov 13

The SkinCast team cuts off temporal leakage through a season-based split, that is, training on data from earlier seasons and evaluating on data from later seasons.

Unverified50%

All Facts (8)

Unverified

SkinCast is the core sales forecasting model in the SkinAdvisor framework, capable of predicting the day-by-day sales trajectory of unreleased skins over 180 days.

50%
Unverified

SkinCast contains four encoders: a pretrained Informer that processes player preference time series; a pretrained Player-Hero-Skin graph embedder that captures relational priors; and two lightweight Transformers that encode skin attributes and release context, respectively.

50%
Unverified

SkinCast's fused features pass through a shared backbone with local-plus-global day-x positional encoding, then through an N-layer Informer decoder that outputs the forecast sequence in a single non-autoregressive pass.

50%
Unverified

Non-autoregressive decoding requires only one forward pass at inference time, significantly reducing latency, and is one of the key designs enabling SkinCast to achieve millisecond-level response on CPU.

50%
Unverified

Offline experiments show that SkinCast reduces RMSE by about 20% compared with the strongest baseline, Informer, and by about 60% compared with classical methods.

50%
Unverified

Stress tests conducted on a 16-core CPU server confirmed that SkinCast's p99 latency is below 8 milliseconds and its peak throughput exceeds 1100 rps.

50%
Unverified

The paper points out that pretrained embeddings may leak temporal information, requiring a season-based split to prevent data leakage.

50%
Unverified

The SkinCast team cuts off temporal leakage through a season-based split, that is, training on data from earlier seasons and evaluating on data from later seasons.

50%

Source Articles