Analysis of machine learning models and technical analysis on SPY ETF
With the rise in computing power and high network speeds, stock trading has become largely algorithm-driven. Researchers and trading houses look to sophisticated quantitative methods to derive profit from the market, and within this machine learning is a rising area of interest. However, there is a...
Saved in:
| 主要作者: | |
|---|---|
| 其他作者: | |
| 格式: | Final Year Project |
| 語言: | English |
| 出版: |
Nanyang Technological University
2023
|
| 主題: | |
| 在線閱讀: | https://hdl.handle.net/10356/165872 |
| 標簽: |
添加標簽
沒有標簽, 成為第一個標記此記錄!
|
| 機構: | Nanyang Technological University |
| 語言: | English |
| 總結: | With the rise in computing power and high network speeds, stock trading has become largely algorithm-driven. Researchers and trading houses look to sophisticated quantitative methods to derive profit from the market, and within this machine learning is a rising area of interest. However, there is a need for more research in comparing the predictive and trading performance of sequential machine learning models and temporal technical indicators. This study investigates the performance of the LSTM, GRU, ARIMA and combined LSTMGRU models as well as the technical indicators 'speed' and 'frequency'. We test our models and indicators on high-frequency tick data of the SPY ETF dataset from 2016 to 2022 and review results for individual years as well as for combined years. Our research shows that the LSTM and ARIMA model have similar capabilities and that temporal technical indicators can improve trading abilities. We also show that the LSTMGRU model improves trading performance from the LSTM and GRU models individually. Academics will be able to use our findings to further improve predictive performance and profitability in trading research. |
|---|