LSTM-MSNet: Leveraging Forecasts on Sets of Related Time Series with Multiple Seasonal Patterns(Paper Summary and Implementation)

Introduction

Architecture of LSTM-MSNet

Layer 1 : Normalisation and Variance Stabilisation Layer

Layer 2: Seasonal Decomposition

Proposed Methods for Decomposition

Prophet

Layer 3 : Recurrent Layer

Moving Window Transformation :

Training Paradigms

Deseasonalised Approach (DS)

LSTM Learning Scheme

Local Normalization

Loss Function

Post-Processing Layer

Error Metrics

Conclusion

References:

Data Science Enthusiast | Advanced Analytics Intern at EY