Predict Stock Prices Using Neural Network and Random Forest Case Study of Stock Bank Mellat

One of the important issues in statistical science is the prediction of nonlinear models. In the present study, using the Perceptron neural network model and the stochastic forest model, the stock price of Bank Mellat has been predicted during ten years between 1990 and 1999. The MAPE criterion has been used as a measurement criterion. Both are explained in the field of supervised learning. Technical indicators such as MACD, SO, OBV, RSI% RW, etc. have been used as independent variables. Experimental findings from a ten-year study show well that both models alone can predict stock prices, but the neural network model performed better than the random forest, so it has better predictive power.