US recession prediction using statistical and natural language processing methods
Bibliographic record
Abstract
This study mainly predicts the recession in the United States. We build our model based on the data of more than ten recessions experienced by the United States since the mid-20th century. Our research can be divided into two parts, one part is a machine learning model constructed using econometrics theory, and the other part is a text analysis model based on natural language processing (NLP) techniques. We collected quarterly data from January 1, 1950, to September 1, 2020, to examine each historical recessionary period. We select key macroeconomic variables such as real GDP growth rate, unemployment rate, and interest rates as variables to build the machine learning model. Depending on the data type and model accuracy, we adopted three models, Support Vector Classification (SVC), Naive Bayes, and Logistic Regression, where the SVC model has the highest accuracy, above 80%. Regarding NLP models, we choose the reports based on Bank of International Settlements central bank speeches (BIS) to complete the relevant analysis. We evaluate bag-of-words and convolutional neural networks in conjunction with Epoch loss to determine how well the model's predictions match the actual data. Although we have debugged the NLP model many times, its accuracy still needs to be higher than that of the econometric model. How to effectively improve the prediction accuracy of the NLP model will be the main problem we hope to solve in the future.
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 | 0.002 | 0.008 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.001 |
| Bibliometrics | 0.003 | 0.002 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.001 | 0.000 |
| 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".