Performance Evaluation of Artificial Neural Network for Usability Assessment of E-commerce Websites

In E-commerce, website is the bask point of contact between business and its customers, hence for online vendors, it is imperative that their B2C website be usable. The present study comprehensively explores the outstanding ability of Artificial Neural Networks (ANNs) to uncover knowledge hidden in data and thus discover relationships between input and output variables and entails questionnaire survey approach to assess usability of B2C websites. The performance of various ANN models has been evaluated by altering different parameters of a neural network. Three metrics viz. MSE, MAE and MAPE have been used to measure the performance of the neural network. The empirical results show that best result is obtained by using tansig purelin transfer function with trainlm training function having three nodes at hidden layer with 90 percent data for training. Further, in assessing the usability of E-commerce websites using ANN it was found that although all dimensions were important, System Quality followed by Trust were most significant.