Affinity-Driven Transfer Learning for Load Forecasting

. 2024 Sep 6;24(17):5802. doi: 10.3390/s24175802

Algorithm 1 Affinity-Driven Transfer Learning

Input: Old grid elements Electricity Demand (OGTS)
Input: New grid element Electricity demand (NGTS)
Output:

h^{*}

I - pretraining:
1: train

h_{i}

models on

\forall i \in {1 \dots m}

dataset from OGTS
II - Transfer Learning:
2:

h^{*} = {min}_{h_{i}, i \in {1 \dots m}} T A S (τ_{i}, t a u_{n e w})

3: Train

h^{*}

for few epochs
4: return

h^{*}