Adam Phillips Undergraduate Dissertation 2017/18

Using Support Vector Machine Classification to Predict Future Stock Prices Based on Previous Financial Data

Supervised by E.Vasilaki

Abstract

This project focuses on stock market prediction using Support Vector Machine Classification. The overall aim of the project is to investigate how well a classification algorithm can predict future stock prices, with the main consideration being how well a model can predict future prices based on both daily and weekly values. Two different experiments are deduced for which the time periods, daily and weekly, to determine which of the periods is best suited to short term trading and which is best suited for short term trading.