Time Series Forecasting with Long Short-Term Memory Neural Networks on the Stock Market
Maresia, Erik (2020)
2020
All rights reserved. This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:amk-202002182562
https://urn.fi/URN:NBN:fi:amk-202002182562
Tiivistelmä
The stock market is notoriously difficult to predict, but there are two schools of thought that make approximation possible. Both the fundamental and technical analysis attempt to provide ways in which a human or a machine can predict the market.
Machine learning is an efficient tool to use for this problem. It offers solutions that fit the scope of the problem, such as long short-term memory neural networks. Combining all the knowledge of economics and machine learning, it is possible to attempt to overcome the efficient-market hypothesis, which states that it is impossible to outperform the market.
The aim of this bachelor’s thesis is to demonstrate how a neural network works and how it may be used to solve a problem such as stock market forecasting. Since it is not known how neural networks find patterns, this thesis will give a theory for stock markets to demonstrate what might be happening inside a neural network when it finds patterns.
Machine learning is an efficient tool to use for this problem. It offers solutions that fit the scope of the problem, such as long short-term memory neural networks. Combining all the knowledge of economics and machine learning, it is possible to attempt to overcome the efficient-market hypothesis, which states that it is impossible to outperform the market.
The aim of this bachelor’s thesis is to demonstrate how a neural network works and how it may be used to solve a problem such as stock market forecasting. Since it is not known how neural networks find patterns, this thesis will give a theory for stock markets to demonstrate what might be happening inside a neural network when it finds patterns.