Exploring Information Categories and Artificial Neural Networks Numerical Algorithms in S&P500 Trend Prediction
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
The purpose of this study is to examine three major issues. First, the authors compare the performance of economic information, technical indicators, historical information, and investor sentiment measures in financial predictions using backpropagation neural networks (BPNN). Granger causality tests are applied to each category of information to select the relevant variables that statistically and significantly affect stock market shifts. Second, the authors investigate the effect of combining all of these four categories of information variables selected by Granger causality test on the prediction accuracy. Third, the effectiveness of different numerical techniques on the accuracy of BPNN is explored. The authors include conjugate gradient algorithms (Fletcher-Reeves update, Polak-Ribiére update, Powell-Beale restart), quasi-Newton (Broyden-Fletcher-Goldfarb-Shanno, BFGS), and the Levenberg-Marquardt (LM) algorithm which is commonly used in the literature. Fourth, the authors compare the performance of the BPNN and support vector machine (SVM) in terms of stock market trend prediction. Their comparative study is applied to S&P500 data to predict its future moves. The out-of-sample forecasting results show that (i) historical values and sentiment measures allow obtaining higher accuracy than economic information and technical indicators, (ii) combining the four categories of information does not help improving the accuracy of the BPNN and SVM, (iii) the LM algorithm is outperformed by Polak-Ribière, Powell-Beale, and Fletcher-Reeves algorithms, and (iv) the BPNN outperforms the SVM except when using sentiment measures as predictive information.
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.
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.012 | 0.007 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.004 |
| Open science | 0.001 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
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 it