Variet Agent
|
d1ddf06e5d
|
feat(model): KAP YTM PD floor integration, expanded 226-var search, ADF fix (AIC->BIC), Model#2 with 6-test diagnostics
- Replace hardcoded DEFAULT_PD_FLOORS with build_complete_pd_floor_table() (KAP bond YTM)
- Fix ADF test: autolag='AIC' -> 'BIC' for small sample (N=26) robustness
- Expand variable search: 40 -> 226 vars (log/diff/return/lag2), 1.9M combos
- Select Model #2: HOUSING_PRICE + CREDIT_SPREAD_LAG1 + CURRENT_ACCOUNT_R
- Add 6-test diagnostics table to AR1 sheet (ADF/LB/DW/BP/ARCH/Shapiro)
- Add Korean variable names for transformed variables
- Generate report v7 with full diagnostics
|
2026-03-12 00:06:23 +09:00 |
|
Variet Agent
|
87725b7c19
|
feat: 3-variable macro model (USDKRW+RETAIL_SALES+INVEST_RATE), forced_vars support, methodology sync
|
2026-03-11 17:30:06 +09:00 |
|
Variet Agent
|
a4c9a702ad
|
feat(pipeline): wire 7x7 Zt + 8x8 CCC expansion into main.py lifetime PD
|
2026-03-11 16:19:58 +09:00 |
|
Variet Agent
|
811d6ee843
|
feat(macro): comprehensive variable exploration, R²=0.028→0.747
- New: data/macro_analysis.py (15 base × 6 transforms = 116 candidates)
- Top correlations: CORP_AA_LOGR(r=-0.75), credit spread, term spread
- Exhaustive 3-var search (1749 combos), best adj.R²=0.71
- Modified: data/macro_data.py
- Added GOVT_3Y, CORP_AA, CORP_BBB ECOS queries + fallback data
- New: compute_derived_features() for optimal 3 predictors
- Modified: main.py
- Computes derived features + passes combined input to stepwise
- Scenario paths now include derived features for prediction
- Selected 3 variables: CORP_AA_LOGR, CPI_GROWTH, CREDIT_SPREAD_LAG1
- All 8/8 validation tests pass (incl. R² now Pass)
|
2026-03-11 06:55:02 +09:00 |
|
Variet Agent
|
8af743e6f3
|
feat(data): parse 3-agency PDF transition matrices to CSV #task-290
- New: data/parse_pdf_matrices.py (KR/NICE/SCI PDF parser)
- KR: text-based parser (space-separated numbers + dashes)
- NICE: text-based parser (clean numeric format)
- SCI: pdfplumber table extraction (column-position-aware)
- WR redistribution, B이하→B mapping, CCC extrapolation from PD patterns
- Modified: data/transition_matrices.py (added source='real' loader)
- Modified: config.yaml (data.transition_source: 'real')
- Modified: main.py (reads transition source from config)
- Output: 112 CSV files (KR/NICE/SCI/AVG × 28 years)
|
2026-03-11 01:07:27 +09:00 |
|
Variet Agent
|
3a9374c61a
|
feat: Lifetime PD (50yr) - Belkin & Suchower + Vasicek model
- Belkin & Suchower (1998) credit cycle index (Zt) estimation via WLS
- Vasicek single-factor conditional PD/TM model
- Macro-Zt OLS regression with stepwise variable selection
- 3-scenario (boom/neutral/recession) 50yr PD projection
- Statistical validation suite (ADF, Ljung-Box, R2, ARCH)
- BOK ECOS API integration with fallback data
- Visualization module (7 chart types)
- Detailed theoretical methodology docs/methodology.md
|
2026-03-10 21:57:34 +09:00 |
|