An optimized ANN Model for predicting the efficiency of perovskite solar cell using MATLAB

The amalgamation of material science genome and algorithmic development has elevated the evolution of material science. Traditional methods of material discovery, development and deployment takes a long time frame. Therefore, machine learning models which primarily learns from past data helps in catering to the inherit limitations of conventional methods used in material science. Hence we demonstrate the potential of deep learning via Artificial Neural Network (ANN) which utilizes radical features to predict the efficiency of perovskite solar cell. Dataset was collected varies technical papers. The trained model then predicts the efficiency on unseen perovskite data. This paper also finds insights of challenges faced with ANN and how it could beimprovised in the near future.