A PREDICTION OF DIABETES POSSIBILITY WITH DEEP LEARNING: LSTM AND ESA HYBRID MODEL

Abstract

The current research uses a deep learning model for the prediction of diabetes cases from Pima Indians database. The model combines Long-short term memory networks (LSTM) and convolutional neural networks (CNN) through an explicit semantic analysis (ESA). The proposed model is applied to the dataset using different settings, where the accuracy of the hybrid model reached to 86.4%. It was the best accuracy when compared to other models using a single classification by either LSTM or CNN.