A Time Series Analysis Of The Toronto-Dominion Bank Stock Price

Abstract

This paper presents a time series analysis of the Toronto-Dominion Bank stock (TD) for 10 months into the future. The data set was collected from Yahoo Finance. The time period of the dataset is from January 2008 to November 2015 and the prediction goes till September 2016. The attribute measured is the value of the stock measured at the beginning of the month. The data was analyzed using the TSA package in the R statistical language. Exploring the data through multiple tests gave five possible models that could represent the trends in the data. The model's predictions were then compared using MPE, MSE, MAE, and MAPE to find the prediction accuracy. The Regression AR(1) model was clearly the best model after looking at these values. This model also did not have insignificant coefficients that were an issue for the other models. The value of a company is the results of many factors both external and internal, despite this, the Regression AR(1) model was still able to give fairly accurate predictions 10 months into the future.

Discipline: Statistics

Faculty Mentor: Dr. Cristina Anton