Algorithmic Trading, Human against AI
(2023) NEKH02 20222Department of Economics
- Abstract
- This thesis examines how the performance of a rule based trading algorithm, the Golden Cross Trading Algorithm, and an Artificial Neural Network constructed algorithm, the Long Short-Term Memory, perform against each other and a buy- and-hold strategy, when trading on the equity market. The underlying assets will be nine large American tech companies, all listed on Nasdaq. In order to distinguish and be able to interpret how each model for each stock performs, several financial measurements will be produced, there among alpha and beta values. Overall, it can be said that Golden Cross points towards having difficulties to overperform the Buy-and-Hold strategy, while Long Short-Term Memory points towards having pos- sibilities to overperform... (More)
- This thesis examines how the performance of a rule based trading algorithm, the Golden Cross Trading Algorithm, and an Artificial Neural Network constructed algorithm, the Long Short-Term Memory, perform against each other and a buy- and-hold strategy, when trading on the equity market. The underlying assets will be nine large American tech companies, all listed on Nasdaq. In order to distinguish and be able to interpret how each model for each stock performs, several financial measurements will be produced, there among alpha and beta values. Overall, it can be said that Golden Cross points towards having difficulties to overperform the Buy-and-Hold strategy, while Long Short-Term Memory points towards having pos- sibilities to overperform the Buy-and-Hold strategy. Both algorithms have almost the same financial measurements in all categories, but big differences can be seen mainly in the number of trades, where Long-Short Term Memory executes approxi- mately 19 trades per year (buy and sell included) compared to Golden Cross’s 2 to 3. For future research, the conclusion can be drawn that more underlying stocks are needed to be tested on, and also over more time periods, in order to come up with more reliable results. (Less)
Please use this url to cite or link to this publication:
http://lup.lub.lu.se/student-papers/record/9113613
- author
- Mild, Gustaf LU
- supervisor
- organization
- course
- NEKH02 20222
- year
- 2023
- type
- M2 - Bachelor Degree
- subject
- keywords
- Algoritmic Trading, Buy-and-Hold, Golden Cross Trading Algorithm, Long Short-Term Memory, NASDAQ, Tech Companies.
- language
- English
- id
- 9113613
- date added to LUP
- 2023-05-29 12:03:18
- date last changed
- 2023-05-29 12:03:18
@misc{9113613, abstract = {{This thesis examines how the performance of a rule based trading algorithm, the Golden Cross Trading Algorithm, and an Artificial Neural Network constructed algorithm, the Long Short-Term Memory, perform against each other and a buy- and-hold strategy, when trading on the equity market. The underlying assets will be nine large American tech companies, all listed on Nasdaq. In order to distinguish and be able to interpret how each model for each stock performs, several financial measurements will be produced, there among alpha and beta values. Overall, it can be said that Golden Cross points towards having difficulties to overperform the Buy-and-Hold strategy, while Long Short-Term Memory points towards having pos- sibilities to overperform the Buy-and-Hold strategy. Both algorithms have almost the same financial measurements in all categories, but big differences can be seen mainly in the number of trades, where Long-Short Term Memory executes approxi- mately 19 trades per year (buy and sell included) compared to Golden Cross’s 2 to 3. For future research, the conclusion can be drawn that more underlying stocks are needed to be tested on, and also over more time periods, in order to come up with more reliable results.}}, author = {{Mild, Gustaf}}, language = {{eng}}, note = {{Student Paper}}, title = {{Algorithmic Trading, Human against AI}}, year = {{2023}}, }