Add Sub-sampling approach for Kernel Alignment

[OkuyanBoga]

• Create a loss function inherited from SVCLoss using data batcher functions as described in [1].
• Add it to existing kernel alignment notebook.

[1] Sahin, M. E., Symons, B. C., Pati, P., Minhas, F., Millar, D., Gabrani, M., ... & Mensa, S. (2024). Efficient Parameter Optimisation for Quantum Kernel Alignment: A Sub-sampling Approach in Variational Training. arXiv preprint arXiv:2401.02879.