FIDELITY STOCK PRICE PREDICTION

OVERVIEW

• In this project we predict the price of Fidelity National Financial Inc.
• In this project we make use of LSTM recurrent neural network to make time series forecasting.

DATASET

• I downloaded this data from yahoo_finance.
• I downloaded the data since 2.08.2016.
• You can access the data i used by downloading the file named Fidelity.csv.

ATTRIBUTES DESCRIPTION

Column	Description
DATE	Date of the which the below details was recorded
OPEN	Value of the last stock that traded on a particular day
HIGH	Highest price the stock traded for on a particular day
LOW	Lowest price the stock traded for on a particular day
CLOSE	Value of the last stock that traded on a particular day
ADJ_CLOSE	Closing price of the stock on a particular day
VOLUME	Amount of stock bought or sold on a particular day

TECHNIQUES USED

• Moving Average
• LSTM model

LSTM ARCHITECTURE

• We make use of two LSTM layers followed by a Dense layer.

MOVING AVERAGE PERFORMANCE

• From this we can see that the prediction using the moving average doesnt seem to work very well.

LSTM MODEL PERFORMANCE

• From this we can see that our LSTM model makes accurate predictions on the stock price.
• Green is the predicted price while Orange is the actual price.
• The predicted and actual price are nearly the same.