Forecasting etfs- price movements using convolutional neural networks - methodology and comparison of industries

Casqueiro, Patrícia XufreEngelsmann, Moritz Peter Gerhard2023-06-092023-06-092022-01-212021-12-17http://hdl.handle.net/10362/153750The aim of this paper is to achieve two goals. Firstly, build and apply a convolutional neural network to make predictions on historical data of the Vanguard Industrials ETF (VIS) in the form of Buy, Hold and Sell signals. Secondly, making comparisons among different indus triesin order to derive potential performance deviations. By using three image encoding tech niques and a randomly generated model for comparison purposes, some promising results have been achieved. Nevertheless, several classic strategies and the market performance could not be beaten, mainly because model predictions for Buy and Sell signals showed weaknesses.engForecastingTradingDeep learningForecasting etfs- price movements using convolutional neural networks - methodology and comparison of industries - focus on industrials etfmaster thesis203245431