Mtanh-Attention-BiLSTM model for prediction of Automobile Export Customer Satisfaction

Abstract

Attention based Bidirectional LSTM is used in machine learning for Data Classification and predictions. The activation function is one of the important aspects which facilitate the deep training by introducing the non-linearity data into the learning & linear process. There are various types of existing activation functions like Sigmoid, tanh, Relu etc. To increase the performance in terms of accuracy the novel research work Mtanh-Attention-BiLSTM model for classification and prediction was proposed. This novel activation functions reduces the gradient problems and transform the nonlinear into linear form. The training and classification was done with the primary export dataset and standard car dataset also. The performance improvement was observed with increase in accuracy compared with the existing model.