High-frequency stock return prediction using state-of-the-art deep learning models
Bibliographic record
Abstract
Determining stock price movements is a challenging problem because stock prices are often influenced by multiple factors such as economic, political, business, and human behavior. In this paper, we will attempt different modeling methods for two types of data, a total of 40 Dow Jones Industrial Index components, to verify the effectiveness of daily and high-frequency data for stock price prediction. Furthermore, we will attempt to validate the performance of LSTM model in stock price prediction, and also try to improve its performance by incorporating an attention mechanism. We assume that adding an attention layer to LSTM model would improve model performance in our data sets, especially in high-frequency data, since the data set would contain a huge amount of noise. Our results indicate that the simple LSTM performs better than the attention-based LSTM for both data types of prediction tasks with a benchmark of the number of stock prediction outcomes that outperform the number of those in other model, which is 24 out 40 stocks, which refutes our initial assumptions and does not validate whether adding attention mechanism is useful for solving the shallow layers and gradient vanishing problem and thus improving the LSTM model performance.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".