{ "cells": [ { "cell_type": "markdown", "id": "b9f72179", "metadata": {}, "source": [ "# 03 결과 도출 — 옵션 비교표·그림, 이전 판 대비 변화, 검증\n", "\n", "`02_분석.ipynb`의 산출물로 4장의 옵션 비교표와 그림을 만든다. 마지막 셀은 다음을 대조한다(반올림 범위 안에서 일치해야 통과).\n", "\n", "- 정비판 그리드(정비판 외생) ↔ 2026-10 정비 실행에서 동결한 기록 그리드(`data/published/grid_summary_*.csv`)와 핵심 수치(`config.PUBLISHED`)\n", "- 이전 판 외생으로 다시 계산한 그리드 ↔ 이전 공표 그리드(`data/published_prev/`, 2026-08-28·09-01 실행)와 논문 1–3 본문 인용값(`config.PREVIOUS`)" ] }, { "cell_type": "code", "execution_count": 1, "id": "e7d94f54", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:40.198522Z", "iopub.status.busy": "2026-10-08T00:05:40.198522Z", "iopub.status.idle": "2026-10-08T00:05:40.900257Z", "shell.execute_reply": "2026-10-08T00:05:40.899753Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Python 3.12.10 · pandas 2.2.3 · numpy 2.4.6 · matplotlib 3.11.0 · BLAS 스레드 1\n", "표본 2000Q1–2025Q4 · train ~2021Q4 · valid ~2023Q4 · test 이후 · 옵션 12개 · P=4, Q=2, λ=0.2\n", " 있음 data\\inputs\\levels_sa.csv\n", " 있음 data\\inputs\\levels_62.csv\n", " 있음 data\\inputs\\levels_69.csv\n", " 있음 data\\inputs\\levels_agg.csv\n", " 있음 data\\inputs\\io_linkage_W.csv\n", " 있음 data\\inputs\\io_linkage_62.csv\n", " 있음 data\\inputs\\io_linkage_69.csv\n", " 있음 data\\inputs\\exog_hist_v2_20260828.csv\n", " 있음 data\\inputs\\exog_hist_v2.csv\n" ] } ], "source": [ "import config as C # 선형대수 스레드를 1개로 고정하므로 numpy보다 먼저 불러온다\n", "import os, hashlib, time\n", "import numpy as np, pandas as pd\n", "import varx_grid as V\n", "pd.set_option(\"display.float_format\", lambda v: f\"{v:,.3f}\")\n", "pd.set_option(\"display.max_columns\", 30)\n", "C.setup()\n", "import matplotlib.pyplot as plt\n", "import koreanize_matplotlib # noqa: F401 (한글 글꼴)\n", "grids = {g: pd.read_csv(os.path.join(C.OUTPUT, f\"grid_{g}.csv\"), encoding=\"utf-8-sig\") for g in C.GRIDS}\n", "prev = {g: pd.read_csv(os.path.join(C.OUTPUT, f\"grid_prev_{g}.csv\"), encoding=\"utf-8-sig\") for g in C.GRIDS}\n", "ORDER = [o[\"옵션\"] for o in V.options()]\n", "M = [\"RMSE_train\", \"RMSE_valid\", \"RMSE_test\"]" ] }, { "cell_type": "markdown", "id": "79d55f60", "metadata": {}, "source": [ "## 표 1. 12개 옵션의 예측오차(RMSE, 100·log 수준) — 분류별 valid·test(정비판)\n", "\n", "62·69분류는 36개 공표 산업을 고정 비중으로 나눈 패널이어서 서로 다른 성장률 계열이 각각 32개·34개뿐이다(01 2절).\n", "따라서 분류 간 RMSE 차이는 산업 동학의 차이보다 산업별 가중의 차이를 많이 반영하며, 비교는 같은 분류 안의 옵션끼리 하는 것이 맞다." ] }, { "cell_type": "code", "execution_count": 2, "id": "cf0de8c8", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:40.901259Z", "iopub.status.busy": "2026-10-08T00:05:40.901259Z", "iopub.status.idle": "2026-10-08T00:05:40.945670Z", "shell.execute_reply": "2026-10-08T00:05:40.945670Z" } }, "outputs": [ { "data": { "text/html": [ "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	36	62	69
	RMSE_valid	RMSE_test	RMSE_valid	RMSE_test	RMSE_valid	RMSE_test
옵션
차분_minnesota_varx	7.296	3.626	6.188	3.332	6.232	3.525
차분_minnesota_var	7.405	3.597	6.355	3.322	6.389	3.581
차분_io_varx	6.744	3.106	6.181	3.306	6.230	3.449
차분_io_var	6.885	2.916	6.316	3.274	6.358	3.443
수준_minnesota_varx	13.939	4.755	13.426	4.539	12.944	5.128
수준_minnesota_var	14.823	5.178	14.118	4.866	13.795	5.279
수준_io_varx	10.325	3.804	12.476	4.485	12.475	4.442
수준_io_var	10.417	3.827	13.566	4.907	13.491	4.735
VECM_minnesota_varx	8.111	4.173	7.600	4.155	8.067	4.963
VECM_minnesota_var	8.387	4.166	7.906	4.286	8.395	5.008
VECM_io_varx	8.660	4.153	7.490	4.091	7.929	4.969
VECM_io_var	8.739	4.196	7.837	4.215	8.344	4.867

\n" ], "text/plain": [ "" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "t1 = pd.concat({g: grids[g].set_index(\"옵션\")[[\"RMSE_valid\", \"RMSE_test\"]] for g in [\"36\", \"62\", \"69\"]}, axis=1).loc[ORDER]\n", "t1.to_csv(os.path.join(C.OUTPUT, \"table1_grid_rmse.csv\"), encoding=\"utf-8-sig\")\n", "t1.style.format(\"{:.3f}\").highlight_min(axis=0, props=\"font-weight:bold\")" ] }, { "cell_type": "markdown", "id": "7c623c88", "metadata": {}, "source": [ "## 표 2. 분류별 선택 옵션(valid 최소)과 시험 구간 성능(정비판)\n", "\n", "선택 옵션과 시험 구간 최소 옵션을 함께 보인다. 총량은 수준 명세가 선택되며 두 사전분포는 같은 모형이다(동률).\n", "5장은 총량의 경로를 BOKDPM 기준선에서 가져오고, 총량 VARX(수준·산업연관표·VARX)는 시나리오 간 차이를 내는 데만 쓴다." ] }, { "cell_type": "code", "execution_count": 3, "id": "64da9c06", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:40.946674Z", "iopub.status.busy": "2026-10-08T00:05:40.946674Z", "iopub.status.idle": "2026-10-08T00:05:40.954654Z", "shell.execute_reply": "2026-10-08T00:05:40.954150Z" } }, "outputs": [ { "data": { "text/html": [ "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	선택 옵션(valid 최소)	RMSE_train	RMSE_valid	RMSE_test	test 최소 옵션	test 최소	이전 판 선택
분류
36	차분_io_varx	3.691	6.744	3.106	차분_io_var	2.916	차분_io_varx
62	차분_io_varx	2.023	6.181	3.306	차분_io_var	3.274	차분_io_varx
69	차분_io_varx	2.098	6.230	3.449	차분_io_var	3.443	차분_io_varx
AGG	수준_minnesota_var = 수준_io_var	0.883	0.414	0.554	수준_minnesota_var	0.554	수준_minnesota_var = 수준_io_var

\n" ], "text/plain": [ "" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "t2 = []\n", "for g, d in grids.items():\n", " best = V.select(d)\n", " r = d.set_index(\"옵션\").loc[best[0]]\n", " tmin = d.loc[d[\"RMSE_test\"].idxmin()]\n", " t2.append({\"분류\": g, \"선택 옵션(valid 최소)\": \" = \".join(best), **{m: r[m] for m in M},\n", " \"test 최소 옵션\": tmin[\"옵션\"], \"test 최소\": tmin[\"RMSE_test\"],\n", " \"이전 판 선택\": \" = \".join(V.select(prev[g]))})\n", "t2 = pd.DataFrame(t2).set_index(\"분류\")\n", "t2.to_csv(os.path.join(C.OUTPUT, \"table2_selected.csv\"), encoding=\"utf-8-sig\")\n", "t2.style.format({m: \"{:.3f}\" for m in M} | {\"test 최소\": \"{:.3f}\"})" ] }, { "cell_type": "markdown", "id": "8d51b04e", "metadata": {}, "source": [ "## 표 3. 축별 평균 RMSE — 명세·사전분포·외생(정비판)\n", "\n", "36·62·69분류 각각에서 해당 축의 수준을 공유하는 옵션(4개 또는 6개)의 평균이다. 총량은 사전분포 축이 의미가 없으므로 뺀다." ] }, { "cell_type": "code", "execution_count": 4, "id": "8be85697", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:40.955653Z", "iopub.status.busy": "2026-10-08T00:05:40.955155Z", "iopub.status.idle": "2026-10-08T00:05:40.966537Z", "shell.execute_reply": "2026-10-08T00:05:40.966537Z" } }, "outputs": [ { "data": { "text/html": [ "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

		valid	test
	분류	36	62	69	36	62	69
축	수준
명세	VECM	8.474	7.708	8.184	4.172	4.187	4.952
수준	12.376	13.397	13.176	4.391	4.699	4.896
차분	7.083	6.260	6.302	3.311	3.308	3.500
사전	io	8.629	8.978	9.138	3.667	4.046	4.317
minnesota	9.994	9.266	9.304	4.249	4.083	4.581
외생	var	9.443	9.350	9.462	3.980	4.145	4.485
varx	9.179	8.894	8.980	3.936	3.985	4.413

\n" ], "text/plain": [ "" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "three = {g: grids[g] for g in [\"36\", \"62\", \"69\"]}\n", "t3 = pd.concat({\"valid\": V.axis_means(three, \"RMSE_valid\"), \"test\": V.axis_means(three, \"RMSE_test\")}, axis=1)\n", "t3.to_csv(os.path.join(C.OUTPUT, \"table3_axis_means.csv\"), encoding=\"utf-8-sig\")\n", "t3.style.format(\"{:.3f}\")" ] }, { "cell_type": "markdown", "id": "69c5ccfb", "metadata": {}, "source": [ "## 그림 1. 69분류 12개 옵션의 구간별 RMSE(valid 오름차순, 정비판)" ] }, { "cell_type": "code", "execution_count": 5, "id": "a757471e", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:40.967540Z", "iopub.status.busy": "2026-10-08T00:05:40.967540Z", "iopub.status.idle": "2026-10-08T00:05:41.145645Z", "shell.execute_reply": "2026-10-08T00:05:41.145139Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "d = grids[\"69\"].sort_values(\"RMSE_valid\")\n", "x = np.arange(len(d))\n", "fig, ax = plt.subplots(figsize=(10, 4.2))\n", "for col, mk, lab in [(\"RMSE_train\", \"o\", \"train\"), (\"RMSE_valid\", \"s\", \"valid\"), (\"RMSE_test\", \"^\", \"test\")]:\n", " ax.plot(x, d[col], mk + \"-\", label=lab)\n", "ax.set_xticks(x); ax.set_xticklabels(d[\"옵션\"], rotation=45, ha=\"right\", fontsize=8)\n", "ax.set_ylabel(\"RMSE(100·log 수준)\"); ax.grid(alpha=.3); ax.legend(frameon=False)\n", "ax.set_title(\"69분류 · 12개 옵션의 1기 앞 예측오차\", loc=\"left\")\n", "fig.tight_layout(); fig.savefig(os.path.join(C.OUTPUT, \"fig1_rmse_69.png\"), dpi=150)" ] }, { "cell_type": "markdown", "id": "4dec5a6d", "metadata": {}, "source": [ "## 그림 2. 분류 × 옵션 valid RMSE(정비판)" ] }, { "cell_type": "code", "execution_count": 6, "id": "0cc8ee9b", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:41.146648Z", "iopub.status.busy": "2026-10-08T00:05:41.146648Z", "iopub.status.idle": "2026-10-08T00:05:41.386929Z", "shell.execute_reply": "2026-10-08T00:05:41.386929Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "

" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "h = pd.DataFrame({g: grids[g].set_index(\"옵션\")[\"RMSE_valid\"] for g in [\"36\", \"62\", \"69\"]}).loc[ORDER]\n", "fig, ax = plt.subplots(figsize=(6.2, 6.4))\n", "im = ax.imshow(h.values, cmap=\"Blues\", aspect=\"auto\")\n", "ax.set_xticks(range(3)); ax.set_xticklabels([f\"{g}분류\" for g in h.columns])\n", "ax.set_yticks(range(len(h))); ax.set_yticklabels(h.index, fontsize=8)\n", "for i in range(h.shape[0]):\n", " for j in range(h.shape[1]):\n", " v = h.values[i, j]\n", " ax.text(j, i, f\"{v:.2f}\", ha=\"center\", va=\"center\", fontsize=8, color=\"white\" if v > np.nanmedian(h.values) * 1.4 else \"black\")\n", "fig.colorbar(im, ax=ax, shrink=.7, label=\"valid RMSE\")\n", "ax.set_title(\"valid RMSE(낮을수록 정확)\", loc=\"left\")\n", "fig.tight_layout(); fig.savefig(os.path.join(C.OUTPUT, \"fig2_valid_heatmap.png\"), dpi=150)" ] }, { "cell_type": "markdown", "id": "015e94ee", "metadata": {}, "source": [ "## 표 4. 점검과 민감도 요약\n", "\n", "02 4–6절의 결과를 모은다. 'A 정비 영향'은 실질금리 정의 변경 전후의 RMSE 차이와 선택이다. 비고의 'io−미네소타'는 차분·VARX에서 두 사전분포의 valid RMSE 차이이며, 음수이면 산업연관표가 더 정확하다." ] }, { "cell_type": "code", "execution_count": 7, "id": "207551a1", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:41.388312Z", "iopub.status.busy": "2026-10-08T00:05:41.386929Z", "iopub.status.idle": "2026-10-08T00:05:41.399307Z", "shell.execute_reply": "2026-10-08T00:05:41.399307Z" } }, "outputs": [ { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	점검	분류	선택	비고
0	A 정비 영향(실질금리 수준→차분)	36	차분_io_varx	이전 판 선택 차분_io_varx, 최대 \|Δvalid\| 0.049, \|Δtest\|...
1	A 정비 영향(실질금리 수준→차분)	62	차분_io_varx	이전 판 선택 차분_io_varx, 최대 \|Δvalid\| 0.013, \|Δtest\|...
2	A 정비 영향(실질금리 수준→차분)	69	차분_io_varx	이전 판 선택 차분_io_varx, 최대 \|Δvalid\| 0.022, \|Δtest\|...
3	A 정비 영향(실질금리 수준→차분)	AGG	수준_minnesota_var = 수준_io_var	이전 판 선택 수준_minnesota_var = 수준_io_var, 최대 \|Δval...
4	B 36 행렬 정규화	36	차분_io_varx	valid 6.983
5	λ=0.05	36	차분_minnesota_varx	valid 6.192, test 3.081, 차분·VARX의 io−미네소타 +0.079
6	λ=0.05	62	차분_minnesota_varx	valid 5.166, test 2.654, 차분·VARX의 io−미네소타 +0.051
7	λ=0.05	69	차분_minnesota_varx	valid 5.314, test 2.849, 차분·VARX의 io−미네소타 +0.086
8	λ=0.05	AGG	수준_minnesota_varx = 수준_io_varx	valid 0.322, test 0.571, 차분·VARX의 io−미네소타 +0.000
9	λ=0.1	36	차분_io_varx	valid 6.384, test 3.119, 차분·VARX의 io−미네소타 -0.009
10	λ=0.1	62	차분_minnesota_varx	valid 5.406, test 2.804, 차분·VARX의 io−미네소타 +0.021
11	λ=0.1	69	차분_minnesota_varx	valid 5.517, test 3.011, 차분·VARX의 io−미네소타 +0.059
12	λ=0.1	AGG	수준_minnesota_varx = 수준_io_varx	valid 0.363, test 0.564, 차분·VARX의 io−미네소타 +0.000
13	λ=0.4	36	차분_io_varx	valid 7.477, test 3.200, 차분·VARX의 io−미네소타 -1.298
14	λ=0.4	62	차분_minnesota_varx	valid 7.374, test 4.241, 차분·VARX의 io−미네소타 +0.170
15	λ=0.4	69	차분_minnesota_varx	valid 7.331, test 4.283, 차분·VARX의 io−미네소타 +0.109
16	λ=0.4	AGG	수준_minnesota_var = 수준_io_var	valid 0.394, test 0.558, 차분·VARX의 io−미네소타 +0.000

\n", "

" ], "text/plain": [ " 점검 분류 선택 \\\n", "0 A 정비 영향(실질금리 수준→차분) 36 차분_io_varx \n", "1 A 정비 영향(실질금리 수준→차분) 62 차분_io_varx \n", "2 A 정비 영향(실질금리 수준→차분) 69 차분_io_varx \n", "3 A 정비 영향(실질금리 수준→차분) AGG 수준_minnesota_var = 수준_io_var \n", "4 B 36 행렬 정규화 36 차분_io_varx \n", "5 λ=0.05 36 차분_minnesota_varx \n", "6 λ=0.05 62 차분_minnesota_varx \n", "7 λ=0.05 69 차분_minnesota_varx \n", "8 λ=0.05 AGG 수준_minnesota_varx = 수준_io_varx \n", "9 λ=0.1 36 차분_io_varx \n", "10 λ=0.1 62 차분_minnesota_varx \n", "11 λ=0.1 69 차분_minnesota_varx \n", "12 λ=0.1 AGG 수준_minnesota_varx = 수준_io_varx \n", "13 λ=0.4 36 차분_io_varx \n", "14 λ=0.4 62 차분_minnesota_varx \n", "15 λ=0.4 69 차분_minnesota_varx \n", "16 λ=0.4 AGG 수준_minnesota_var = 수준_io_var \n", "\n", " 비고 \n", "0 이전 판 선택 차분_io_varx, 최대 |Δvalid| 0.049, |Δtest|... \n", "1 이전 판 선택 차분_io_varx, 최대 |Δvalid| 0.013, |Δtest|... \n", "2 이전 판 선택 차분_io_varx, 최대 |Δvalid| 0.022, |Δtest|... \n", "3 이전 판 선택 수준_minnesota_var = 수준_io_var, 최대 |Δval... \n", "4 valid 6.983 \n", "5 valid 6.192, test 3.081, 차분·VARX의 io−미네소타 +0.079 \n", "6 valid 5.166, test 2.654, 차분·VARX의 io−미네소타 +0.051 \n", "7 valid 5.314, test 2.849, 차분·VARX의 io−미네소타 +0.086 \n", "8 valid 0.322, test 0.571, 차분·VARX의 io−미네소타 +0.000 \n", "9 valid 6.384, test 3.119, 차분·VARX의 io−미네소타 -0.009 \n", "10 valid 5.406, test 2.804, 차분·VARX의 io−미네소타 +0.021 \n", "11 valid 5.517, test 3.011, 차분·VARX의 io−미네소타 +0.059 \n", "12 valid 0.363, test 0.564, 차분·VARX의 io−미네소타 +0.000 \n", "13 valid 7.477, test 3.200, 차분·VARX의 io−미네소타 -1.298 \n", "14 valid 7.374, test 4.241, 차분·VARX의 io−미네소타 +0.170 \n", "15 valid 7.331, test 4.283, 차분·VARX의 io−미네소타 +0.109 \n", "16 valid 0.394, test 0.558, 차분·VARX의 io−미네소타 +0.000 " ] }, "execution_count": 7, "metadata": {}, "output_type": "execute_result" } ], "source": [ "ex = pd.read_csv(os.path.join(C.OUTPUT, \"check_prev_vs_new_summary.csv\"), index_col=0, encoding=\"utf-8-sig\")\n", "g36n = pd.read_csv(os.path.join(C.OUTPUT, \"check_io36_normalized.csv\"), encoding=\"utf-8-sig\")\n", "lam = pd.read_csv(os.path.join(C.OUTPUT, \"lambda_sensitivity.csv\"), encoding=\"utf-8-sig\")\n", "t4 = [{\"점검\": \"A 정비 영향(실질금리 수준→차분)\", \"분류\": g, \"선택\": ex.loc[g, \"선택(정비판)\"],\n", " \"비고\": f\"이전 판 선택 {ex.loc[g, '선택(이전 판)']}, 최대 |Δvalid| {ex.loc[g, '최대 |Δvalid|']:.3f}, |Δtest| {ex.loc[g, '최대 |Δtest|']:.3f}\"} for g in ex.index]\n", "t4.append({\"점검\": \"B 36 행렬 정규화\", \"분류\": \"36\", \"선택\": \" = \".join(V.select(g36n)),\n", " \"비고\": f\"valid {g36n['RMSE_valid'].min():.3f}\"})\n", "for _, r in lam[lam[\"λ\"] != C.LAM].iterrows():\n", " t4.append({\"점검\": f\"λ={r['λ']}\", \"분류\": str(r[\"분류\"]), \"선택\": r[\"valid 최소 옵션\"],\n", " \"비고\": f\"valid {r['RMSE_valid']:.3f}, test {r['RMSE_test']:.3f}, 차분·VARX의 io−미네소타 {r['io − 미네소타']:+.3f}\"})\n", "t4 = pd.DataFrame(t4)\n", "t4.to_csv(os.path.join(C.OUTPUT, \"table4_checks.csv\"), index=False, encoding=\"utf-8-sig\")\n", "t4" ] }, { "cell_type": "markdown", "id": "de182d28", "metadata": {}, "source": [ "## 표 5. 이전 판 대비 변화 — 선택 옵션과 VARX 옵션의 RMSE\n", "\n", "실질금리 정의가 바뀌면 외생을 쓰는 VARX·VECM-X 옵션의 RMSE만 달라진다. 아래 표는 분류별 선택 옵션과 주요 VARX 옵션의 valid·test RMSE를 두 판에서 비교한다." ] }, { "cell_type": "code", "execution_count": 8, "id": "e622a2d2", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:41.400810Z", "iopub.status.busy": "2026-10-08T00:05:41.400313Z", "iopub.status.idle": "2026-10-08T00:05:41.408545Z", "shell.execute_reply": "2026-10-08T00:05:41.408545Z" } }, "outputs": [ { "data": { "text/html": [ "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

		valid(정비판)	valid(이전 판)	test(정비판)	test(이전 판)
분류	옵션
36	수준_io_varx	10.325	10.327	3.804	3.859
차분_io_varx	6.744	6.793	3.106	3.093
62	수준_io_varx	12.476	12.479	4.485	4.569
차분_io_varx	6.181	6.179	3.306	3.293
69	수준_io_varx	12.475	12.476	4.442	4.488
차분_io_varx	6.230	6.235	3.449	3.430
AGG	수준_io_varx	0.419	0.417	0.554	0.543
수준_minnesota_var	0.414	0.414	0.554	0.554
차분_io_varx	0.489	0.425	0.632	0.586

\n" ], "text/plain": [ "" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "rows = []\n", "for g in C.GRIDS:\n", " n, p_ = grids[g].set_index(\"옵션\"), prev[g].set_index(\"옵션\")\n", " for o in sorted(set(V.select(grids[g])[:1] + [\"차분_io_varx\", \"수준_io_varx\"])):\n", " rows.append({\"분류\": g, \"옵션\": o, \"valid(정비판)\": n.loc[o, \"RMSE_valid\"], \"valid(이전 판)\": p_.loc[o, \"RMSE_valid\"],\n", " \"test(정비판)\": n.loc[o, \"RMSE_test\"], \"test(이전 판)\": p_.loc[o, \"RMSE_test\"]})\n", "t5 = pd.DataFrame(rows).set_index([\"분류\", \"옵션\"])\n", "t5.to_csv(os.path.join(C.OUTPUT, \"table5_vs_prev.csv\"), encoding=\"utf-8-sig\")\n", "t5.style.format(\"{:.3f}\")" ] }, { "cell_type": "markdown", "id": "8fe1cc11", "metadata": {}, "source": [ "## 표 6. 실질금리 계수의 부호(02 4-1절)" ] }, { "cell_type": "code", "execution_count": 9, "id": "1c5658fd", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:41.409549Z", "iopub.status.busy": "2026-10-08T00:05:41.409549Z", "iopub.status.idle": "2026-10-08T00:05:41.416061Z", "shell.execute_reply": "2026-10-08T00:05:41.416061Z" } }, "outputs": [ { "data": { "text/html": [ "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

		산업 수	동시항 음수 비율(%)	합계 음수 비율(%)	가중평균 동시항	가중평균 합계
분류	판
36	정비판(1차 차분)	36	52.8	50.0	+0.0426	+0.1192
이전 판(수준)	36	27.8	19.4	+0.1478	+0.2126
62	정비판(1차 차분)	62	43.5	45.2	+0.0136	+0.0148
이전 판(수준)	62	25.8	19.4	+0.1008	+0.1948
69	정비판(1차 차분)	69	47.8	44.9	+0.0127	+0.0096
이전 판(수준)	69	27.5	24.6	+0.0995	+0.1979
AGG(수준·io·varx)	정비판(1차 차분)	1	100.0	0.0	-0.1257	+0.0651
이전 판(수준)	1	0.0	0.0	+0.0245	+0.1401

\n" ], "text/plain": [ "" ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "t6 = pd.read_csv(os.path.join(C.OUTPUT, \"real_rate_sign.csv\"), index_col=[0, 1], encoding=\"utf-8-sig\")\n", "t6.to_csv(os.path.join(C.OUTPUT, \"table6_real_rate_sign.csv\"), encoding=\"utf-8-sig\")\n", "t6.style.format({c: \"{:.1f}\" for c in t6.columns if \"비율\" in c} | {c: \"{:+.4f}\" for c in t6.columns if \"가중\" in c})" ] }, { "cell_type": "markdown", "id": "a5518b87", "metadata": {}, "source": [ "## 검증 — 정비판 기록 그리드, 이전 판 공표 그리드\n", "\n", "- 그리드 RMSE는 소수 셋째 자리로 반올림해 기록하므로 허용오차는 0.0005이다(판마다 4개 분류 × 12개 옵션 × 3개 구간 = 144항목).\n", "- 선택 옵션은 valid RMSE(소수 셋째 자리)가 가장 작은 옵션이다. 총량처럼 같은 모형이 동률이면 기록 선택이 동률 집합에 들어 있는지 확인한다.\n", "- 핵심 수치(선택 옵션의 valid·test)는 소수 둘째 자리이므로 허용오차는 0.005이다. 이전 판은 논문 1–3 본문 인용값과 대조한다." ] }, { "cell_type": "code", "execution_count": 10, "id": "913c146c", "metadata": { "execution": { "iopub.execute_input": "2026-10-08T00:05:41.417062Z", "iopub.status.busy": "2026-10-08T00:05:41.417062Z", "iopub.status.idle": "2026-10-08T00:05:41.443264Z", "shell.execute_reply": "2026-10-08T00:05:41.443264Z" } }, "outputs": [ { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

		항목수	최대차이	통과
판	분류
이전 판	36	36	0.000	True
62	36	0.000	True
69	36	0.000	True
AGG	36	0.000	True
정비판	36	36	0.000	True
62	36	0.000	True
69	36	0.000	True
AGG	36	0.000	True

\n", "

" ], "text/plain": [ " 항목수 최대차이 통과\n", "판 분류 \n", "이전 판 36 36 0.000 True\n", " 62 36 0.000 True\n", " 69 36 0.000 True\n", " AGG 36 0.000 True\n", "정비판 36 36 0.000 True\n", " 62 36 0.000 True\n", " 69 36 0.000 True\n", " AGG 36 0.000 True" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "

\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "

	판	항목	계산	기록	허용오차	통과
36	정비판	36 선택 옵션(valid 최소)	차분_io_varx	차분_io_varx	동률 포함	True
73	정비판	62 선택 옵션(valid 최소)	차분_io_varx	차분_io_varx	동률 포함	True
110	정비판	69 선택 옵션(valid 최소)	차분_io_varx	차분_io_varx	동률 포함	True
147	정비판	AGG 선택 옵션(valid 최소)	수준_minnesota_var = 수준_io_var	수준_io_var	동률 포함	True
184	이전 판	36 선택 옵션(valid 최소)	차분_io_varx	차분_io_varx	동률 포함	True
221	이전 판	62 선택 옵션(valid 최소)	차분_io_varx	차분_io_varx	동률 포함	True
258	이전 판	69 선택 옵션(valid 최소)	차분_io_varx	차분_io_varx	동률 포함	True
295	이전 판	AGG 선택 옵션(valid 최소)	수준_minnesota_var = 수준_io_var	수준_io_var	동률 포함	True
296	정비판	정비(36분류) 차분·io·varx valid	6.744	6.740	0.005	True
297	정비판	정비(36분류) 차분·io·varx test	3.106	3.110	0.005	True
298	이전 판	논문(36분류) 차분·io·varx valid	6.793	6.790	0.005	True
299	이전 판	논문(36분류) 차분·io·varx test	3.093	3.090	0.005	True
300	정비판	정비(62분류) 차분·io·varx valid	6.181	6.180	0.005	True
301	정비판	정비(62분류) 차분·io·varx test	3.306	3.310	0.005	True
302	이전 판	논문(62분류) 차분·io·varx valid	6.179	6.180	0.005	True
303	이전 판	논문(62분류) 차분·io·varx test	3.293	3.290	0.005	True
304	정비판	정비(69분류) 차분·io·varx valid	6.230	6.230	0.005	True
305	정비판	정비(69분류) 차분·io·varx test	3.449	3.450	0.005	True
306	이전 판	논문(69분류) 차분·io·varx valid	6.235	6.240	0.005	True
307	이전 판	논문(69분류) 차분·io·varx test	3.430	3.430	0.005	True

\n", "

" ], "text/plain": [ " 판 항목 계산 \\\n", "36 정비판 36 선택 옵션(valid 최소) 차분_io_varx \n", "73 정비판 62 선택 옵션(valid 최소) 차분_io_varx \n", "110 정비판 69 선택 옵션(valid 최소) 차분_io_varx \n", "147 정비판 AGG 선택 옵션(valid 최소) 수준_minnesota_var = 수준_io_var \n", "184 이전 판 36 선택 옵션(valid 최소) 차분_io_varx \n", "221 이전 판 62 선택 옵션(valid 최소) 차분_io_varx \n", "258 이전 판 69 선택 옵션(valid 최소) 차분_io_varx \n", "295 이전 판 AGG 선택 옵션(valid 최소) 수준_minnesota_var = 수준_io_var \n", "296 정비판 정비(36분류) 차분·io·varx valid 6.744 \n", "297 정비판 정비(36분류) 차분·io·varx test 3.106 \n", "298 이전 판 논문(36분류) 차분·io·varx valid 6.793 \n", "299 이전 판 논문(36분류) 차분·io·varx test 3.093 \n", "300 정비판 정비(62분류) 차분·io·varx valid 6.181 \n", "301 정비판 정비(62분류) 차분·io·varx test 3.306 \n", "302 이전 판 논문(62분류) 차분·io·varx valid 6.179 \n", "303 이전 판 논문(62분류) 차분·io·varx test 3.293 \n", "304 정비판 정비(69분류) 차분·io·varx valid 6.230 \n", "305 정비판 정비(69분류) 차분·io·varx test 3.449 \n", "306 이전 판 논문(69분류) 차분·io·varx valid 6.235 \n", "307 이전 판 논문(69분류) 차분·io·varx test 3.430 \n", "\n", " 기록 허용오차 통과 \n", "36 차분_io_varx 동률 포함 True \n", "73 차분_io_varx 동률 포함 True \n", "110 차분_io_varx 동률 포함 True \n", "147 수준_io_var 동률 포함 True \n", "184 차분_io_varx 동률 포함 True \n", "221 차분_io_varx 동률 포함 True \n", "258 차분_io_varx 동률 포함 True \n", "295 수준_io_var 동률 포함 True \n", "296 6.740 0.005 True \n", "297 3.110 0.005 True \n", "298 6.790 0.005 True \n", "299 3.090 0.005 True \n", "300 6.180 0.005 True \n", "301 3.310 0.005 True \n", "302 6.180 0.005 True \n", "303 3.290 0.005 True \n", "304 6.230 0.005 True \n", "305 3.450 0.005 True \n", "306 6.240 0.005 True \n", "307 3.430 0.005 True " ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "검증 통과: 정비판·이전 판 그리드 각 144항목, 선택 옵션 각 4항목, 핵심 수치 각 6항목, 모두 308항목.\n" ] } ], "source": [ "rows = []\n", "def grid_rows(label, calc_grids, pub_paths, sel):\n", " for g in C.GRIDS:\n", " pub = pd.read_csv(pub_paths[g], encoding=\"utf-8-sig\").set_index(\"옵션\")\n", " cal = calc_grids[g].set_index(\"옵션\")\n", " assert set(pub.index) == set(cal.index) == set(ORDER)\n", " for o in ORDER:\n", " for m in M:\n", " rows.append({\"판\": label, \"항목\": f\"{g} {o} {m}\", \"계산\": cal.loc[o, m], \"기록\": pub.loc[o, m], \"허용오차\": 0.0005})\n", " assert pub.index[0] == sel[f\"선택_{g}\"], \"기록 그리드 첫 행과 config의 선택 옵션이 다르다\"\n", " rows.append({\"판\": label, \"항목\": f\"{g} 선택 옵션(valid 최소)\", \"계산\": \" = \".join(V.select(calc_grids[g])),\n", " \"기록\": sel[f\"선택_{g}\"], \"허용오차\": \"동률 포함\"})\n", "grid_rows(\"정비판\", grids, C.PUBLISHED_GRID, C.PUBLISHED)\n", "grid_rows(\"이전 판\", prev, C.PUBLISHED_GRID_PREV, C.PREVIOUS)\n", "for g in [\"36\", \"62\", \"69\"]:\n", " for label, src, P, pre in [(\"정비판\", grids, C.PUBLISHED, \"정비\"), (\"이전 판\", prev, C.PREVIOUS, \"논문\")]:\n", " cal = src[g].set_index(\"옵션\").loc[\"차분_io_varx\"]\n", " for m, key in [(\"RMSE_valid\", \"valid\"), (\"RMSE_test\", \"test\")]:\n", " rows.append({\"판\": label, \"항목\": f\"{pre}({g}분류) 차분·io·varx {key}\", \"계산\": cal[m], \"기록\": P[f\"{pre}_{g}_{key}\"], \"허용오차\": 0.005})\n", "\n", "def passed(r):\n", " if r[\"허용오차\"] == \"동률 포함\":\n", " return r[\"기록\"] in r[\"계산\"].split(\" = \")\n", " return abs(r[\"계산\"] - r[\"기록\"]) <= r[\"허용오차\"] + 1e-9\n", "\n", "res = pd.DataFrame(rows)\n", "res[\"통과\"] = res.apply(passed, axis=1)\n", "res.to_csv(os.path.join(C.OUTPUT, \"verification.csv\"), index=False, encoding=\"utf-8-sig\")\n", "num = res[res[\"항목\"].str.contains(\"RMSE_\")].copy()\n", "num[\"분류\"] = num[\"항목\"].str.split(\" \").str[0]\n", "summary = num.assign(차이=lambda x: (x[\"계산\"].astype(float) - x[\"기록\"].astype(float)).abs()) \\\n", " .groupby([\"판\", \"분류\"]).agg(항목수=(\"통과\", \"size\"), 최대차이=(\"차이\", \"max\"), 통과=(\"통과\", \"all\"))\n", "display(summary)\n", "display(res[~res[\"항목\"].str.contains(\"RMSE_\")])\n", "assert res[\"통과\"].all(), \"기록 수치와 계산 결과가 다르다 — output/verification.csv에서 통과=False 항목을 확인한다\"\n", "print(f\"검증 통과: 정비판·이전 판 그리드 각 144항목, 선택 옵션 각 {len(C.GRIDS)}항목, 핵심 수치 각 6항목, 모두 {len(res)}항목.\")" ] } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.12.10" } }, "nbformat": 4, "nbformat_minor": 5 }