Aim: To detect the fraud in E-Commerce Platform based on Novel Neural Networks and Random Forest Algorithms. Materials and Methods: The performance analysis for maximum accuracy in Fraud detection using Neural Network (N=10) over Random Forest Algorithm which identifies fraud in E-Commerce Platform. GPower is used to compute sample size using a pretest power of 0.8 and an alpha of 0.05. Result: Mean accuracy of Novel Neural Networks 94.54% is high compared to Random Forest 93.21%. Significance value for accuracy and loss is 0.421 (p>0.05). Conclusion: When compared to Random Forest, the accuracy of Novel Neural Networks is higher.