Sparse code multiple access (SCMA) has attracted growing research interests and is a promising technique in order to meet the requirements for future generation wireless communication systems. For the joint combination of Polar Codes (PC) and amplified SCMA signals in our system model, it is essential to evaluate the effects of distortions due to high amplifiers (HPAs) nonlinearities. In this paper, we present the results of the comparative analysis between performance and complexity of Joint PC-SCMA in presence of HPAs. Compared to [1], it is proved that our neural networks based distortion compensation for Joint PC-SCMA system shows a slight performance improvement despite the HPA nonlinearities effects. Also, we demonstrate that our system model offers a good balanced trade-off between BER performance and complexity.