This First Approach involves manually creating sliding windows after normalization, without relying on the Conv1D kernel. This configuration did not include dropout layers.
Module 1 (Sliding Window of Size 1)
Module 1 (Sliding Window of Size 3)
Module 1 (Sliding Window of Size 5)
Module 1 (Sliding Window of Size 7)
Module 1 (Sliding Window of Size 10)
Module 1 (Sliding Window of Size 15)
Module 1 (Best Results)