Stock Prediction Using a Hidden Markov Model Versus a Long Short-Term Memory

de Wit, Bram (2019) Stock Prediction Using a Hidden Markov Model Versus a Long Short-Term Memory. Bachelor's Thesis, Artificial Intelligence.

Preview	Text AI_BA_2019_Bram.pdf Download (1MB) | Preview
	Text toestemming.pdf Restricted to Registered users only Download (139kB)

Abstract

This study will compare the performance of a Hidden Markov Model (HMM) and a Long Short-Term Memory neural network (LSTM) in their ability to predict historical AAPL stock prices. Approximately one hundred other stocks will be used as context vectors in order to predict the following price. This problem is a typical time-series problem, where the models will try to predict the next time step. The performance of the two models will be compared using root mean squared error (RMSE), correlation, mean absolute percentage error (MAPE) and a comparison of the fractal dimension of the predicted stock price sequence. Using k-fold cross validation, for both models the best parameters were chosen. The performance of the best performing models is compared. The results showed that the HMM had a higher performance compared to the LSTM with a RMSE of 2.49 and a MAPE of 4.72, as well as a better fitting fractal dimension when compared to the fractal dimension of the actual data.

Item Type:	Thesis (Bachelor's Thesis)
Supervisor name:	Schomaker, L.R.B.
Degree programme:	Artificial Intelligence
Thesis type:	Bachelor's Thesis
Language:	English
Date Deposited:	15 Aug 2019
Last Modified:	19 Aug 2019 09:18
URI:	https://fse.studenttheses.ub.rug.nl/id/eprint/20683

Actions (login required)

View Item