{ "cells": [ { "cell_type": "markdown", "id": "cell00", "metadata": {}, "source": [ "# Stage One: ML baselines for CA prediction\n", "\n", "**Goal.** Establish a full tabular model suite on the same prediction task later asked of LLMs:\n", "\n", "> Given participant characteristics (demographics → employment → geography → transit), predict **group** and **interpersonal** communication-apprehension subscale scores on the PRCA 6–30 scale.\n", "\n", "These baselines give a non-LLM reference for mean absolute error (MAE). Later LLM persona runs can be compared against the same tiered feature sets and the same ground-truth scores.\n", "\n", "| Piece | Detail |\n", "|---|---|\n", "| Models | Ridge, Elastic Net, k-NN, Random Forest, HistGradientBoosting, XGBoost, MLP |\n", "| Targets | `gt_group_ca`, `gt_interpersonal_ca` |\n", "| Tiers | `demos`, `employment`, `geo`, `transit` (cumulative, matching LLM personas) |\n", "| Validation | Leave-one-out when $N < 8$; otherwise 5-fold CV |\n", "| Primary metric | MAE (aligned with LLM absolute error) |\n", "| CLI | `ca-personas ml-baseline`\n" ] }, { "cell_type": "markdown", "id": "cell01", "metadata": {}, "source": [ "## 1. Setup\n" ] }, { "cell_type": "code", "execution_count": 1, "id": "cell02", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:09.706865Z", "iopub.status.busy": "2026-08-05T22:41:09.706672Z", "iopub.status.idle": "2026-08-05T22:41:11.274456Z", "shell.execute_reply": "2026-08-05T22:41:11.273550Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Project root: /Users/morningstar/Desktop/Cold_Storage/psych755-jjb\n", "Data source: ../sibling_data File A/B/C\n", "Prolific: ['/Users/morningstar/Desktop/Cold_Storage/sibling_data/PRCAProlificExport_FileA.csv', '/Users/morningstar/Desktop/Cold_Storage/sibling_data/PRCAProlificExport_FileB.csv']\n", "Qualtrics: /Users/morningstar/Desktop/Cold_Storage/sibling_data/PRCAQualtricsExport_FileC.csv\n", "Model suite: ['ridge', 'elastic_net', 'knn', 'random_forest', 'hist_gradient_boosting', 'xgboost', 'mlp']\n", "Research tiers (ML baselines; excludes free-text full): ('demos', 'employment', 'geo', 'transit')\n" ] } ], "source": [ "from __future__ import annotations\n", "\n", "import sys\n", "from pathlib import Path\n", "\n", "import matplotlib.pyplot as plt\n", "import pandas as pd\n", "\n", "# Support running from repo root or from notebooks/\n", "ROOT = Path.cwd()\n", "if not (ROOT / \"src\").exists() and (ROOT.parent / \"src\").exists():\n", " ROOT = ROOT.parent\n", "\n", "sys.path.insert(0, str(ROOT / \"src\"))\n", "\n", "from ca_personas.ml_baseline import (\n", " DEFAULT_MODEL_SUITE,\n", " MODEL_LABELS,\n", " TIER_FEATURES,\n", " leaderboard,\n", " mae_pivot,\n", " metrics_wide,\n", " prepare_modeling_frame,\n", " run_stage_one_baselines,\n", " save_baseline_artifacts,\n", ")\n", "from ca_personas.paths import cohort_source_label, full_cohort_paths\n", "from ca_personas.personas import RESEARCH_TIERS\n", "\n", "PROLIFIC, QUALTRICS = full_cohort_paths()\n", "OUT_DIR = ROOT / \"outputs\" / \"ml_baseline\"\n", "FIG_DIR = ROOT / \"docs\" / \"figures\"\n", "OUT_DIR.mkdir(parents=True, exist_ok=True)\n", "FIG_DIR.mkdir(parents=True, exist_ok=True)\n", "\n", "pd.set_option(\"display.max_columns\", 50)\n", "pd.set_option(\"display.float_format\", lambda x: f\"{x:0.3f}\")\n", "print(\"Project root:\", ROOT)\n", "print(\"Data source:\", cohort_source_label())\n", "print(\"Prolific:\", [str(p) for p in PROLIFIC])\n", "print(\"Qualtrics:\", QUALTRICS)\n", "print(\"Model suite:\", list(DEFAULT_MODEL_SUITE))\n", "print(\"Research tiers (ML baselines; excludes free-text full):\", RESEARCH_TIERS)\n" ] }, { "cell_type": "markdown", "id": "cell03", "metadata": {}, "source": [ "## 2. Load joined participants and inspect modeling coverage\n", "\n", "We reuse the project loader so ground-truth PRCA scoring matches the LLM evaluation path (Likert → 1–5, reverse-coded comfort items, subscale sums 6–30).\n" ] }, { "cell_type": "code", "execution_count": 2, "id": "cell04", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:11.276274Z", "iopub.status.busy": "2026-08-05T22:41:11.276101Z", "iopub.status.idle": "2026-08-05T22:41:32.012296Z", "shell.execute_reply": "2026-08-05T22:41:32.011960Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Joined participants with complete CA targets: 241\n", "Models evaluated: ['elastic_net', 'hist_gradient_boosting', 'knn', 'mlp', 'random_forest', 'ridge', 'xgboost']\n", "N= 241 source columns present: ['participant_id', 'Age', 'Sex', 'Country of residence', 'Student status', 'Employment status', 'LocationLatitude', 'LocationLongitude', 'gt_group_ca', 'gt_interpersonal_ca']\n" ] }, { "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", "
participant_idAgeSexCountry of residenceStudent statusEmployment statusgt_group_cagt_interpersonal_ca
0ebab12e9878344fdc93a8c1cad27ff2f709e12b90f0496...36.000MaleUnited StatesYesPart-Time1416
1b6954799ca47bc836ac02ab86e4710c3b0e096d7ace449...41.000MaleUnited StatesNoOther1717
26ac07c3845b0ce5cac4ab3569ea70236b7cfa8fd51576e...39.000MaleUnited StatesNoFull-Time1112
3c67b3266ecd4af12979c1f67d1efffd0d3e9bfbc567d21...61.000MaleUnited KingdomNoOther106
41cefb8d34b92489e8ffd9cf34c469964f0527b1cae33e0...45.000MaleUnited StatesNoFull-Time812
\n", "
" ], "text/plain": [ " participant_id Age Sex \\\n", "0 ebab12e9878344fdc93a8c1cad27ff2f709e12b90f0496... 36.000 Male \n", "1 b6954799ca47bc836ac02ab86e4710c3b0e096d7ace449... 41.000 Male \n", "2 6ac07c3845b0ce5cac4ab3569ea70236b7cfa8fd51576e... 39.000 Male \n", "3 c67b3266ecd4af12979c1f67d1efffd0d3e9bfbc567d21... 61.000 Male \n", "4 1cefb8d34b92489e8ffd9cf34c469964f0527b1cae33e0... 45.000 Male \n", "\n", " Country of residence Student status Employment status gt_group_ca \\\n", "0 United States Yes Part-Time 14 \n", "1 United States No Other 17 \n", "2 United States No Full-Time 11 \n", "3 United Kingdom No Other 10 \n", "4 United States No Full-Time 8 \n", "\n", " gt_interpersonal_ca \n", "0 16 \n", "1 17 \n", "2 12 \n", "3 6 \n", "4 12 " ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "participants, predictions, metrics = run_stage_one_baselines(\n", " PROLIFIC,\n", " QUALTRICS,\n", " tiers=RESEARCH_TIERS,\n", " join_how=\"inner\",\n", " n_neighbors=3,\n", " random_state=42,\n", ")\n", "print(\"Joined participants with complete CA targets:\", len(participants))\n", "print(\"Models evaluated:\", sorted(metrics[\"model\"].unique().tolist()))\n", "print(\n", " \"N=\",\n", " len(participants),\n", " \"source columns present:\",\n", " [\n", " c\n", " for c in [\n", " \"participant_id\",\n", " \"Age\",\n", " \"Sex\",\n", " \"Country of residence\",\n", " \"Student status\",\n", " \"Employment status\",\n", " \"LocationLatitude\",\n", " \"LocationLongitude\",\n", " \"gt_group_ca\",\n", " \"gt_interpersonal_ca\",\n", " ]\n", " if c in participants.columns\n", " ],\n", ")\n", "participants[\n", " [\n", " c\n", " for c in [\n", " \"participant_id\",\n", " \"Age\",\n", " \"Sex\",\n", " \"Country of residence\",\n", " \"Student status\",\n", " \"Employment status\",\n", " \"gt_group_ca\",\n", " \"gt_interpersonal_ca\",\n", " ]\n", " if c in participants.columns\n", " ]\n", "].head()\n" ] }, { "cell_type": "code", "execution_count": 3, "id": "cell05", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:32.013695Z", "iopub.status.busy": "2026-08-05T22:41:32.013617Z", "iopub.status.idle": "2026-08-05T22:41:32.019365Z", "shell.execute_reply": "2026-08-05T22:41:32.019014Z" } }, "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", "
tiern_rowsn_featuresfeatures
0demos2414Age, Sex, Country of residence, Student status
1employment2415Age, Sex, Country of residence, Student status...
2geo2417Age, Sex, Country of residence, Student status...
3transit24113Age, Sex, Country of residence, Student status...
\n", "
" ], "text/plain": [ " tier n_rows n_features \\\n", "0 demos 241 4 \n", "1 employment 241 5 \n", "2 geo 241 7 \n", "3 transit 241 13 \n", "\n", " features \n", "0 Age, Sex, Country of residence, Student status \n", "1 Age, Sex, Country of residence, Student status... \n", "2 Age, Sex, Country of residence, Student status... \n", "3 Age, Sex, Country of residence, Student status... " ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "coverage = []\n", "for tier in RESEARCH_TIERS:\n", " frame = prepare_modeling_frame(participants, tier=tier)\n", " coverage.append(\n", " {\n", " \"tier\": tier,\n", " \"n_rows\": len(frame),\n", " \"n_features\": len([c for c in TIER_FEATURES[tier] if c in frame.columns]),\n", " \"features\": \", \".join([c for c in TIER_FEATURES[tier] if c in frame.columns]),\n", " }\n", " )\n", "pd.DataFrame(coverage)\n" ] }, { "cell_type": "markdown", "id": "cell06", "metadata": {}, "source": [ "## 3. Evaluate the full ML suite\n", "\n", "Cross-validated MAE / RMSE / R² / band accuracy for each model × tier × target.\n" ] }, { "cell_type": "code", "execution_count": 4, "id": "cell07", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:32.020459Z", "iopub.status.busy": "2026-08-05T22:41:32.020389Z", "iopub.status.idle": "2026-08-05T22:41:32.027411Z", "shell.execute_reply": "2026-08-05T22:41:32.027017Z" } }, "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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
tiermodel_labeltargetn_samplesn_featuresmaermser2band_acc
2demosElastic Netgt_group_ca24144.9366.106-0.0290.311
3demosElastic Netgt_interpersonal_ca24144.6485.865-0.0200.361
16employmentElastic Netgt_group_ca24154.6465.8690.0500.456
17employmentElastic Netgt_interpersonal_ca24154.3965.6150.0650.440
30geoElastic Netgt_group_ca24174.6945.8880.0430.465
31geoElastic Netgt_interpersonal_ca24174.3705.5640.0810.456
44transitElastic Netgt_group_ca241134.5065.5740.1430.469
45transitElastic Netgt_interpersonal_ca241134.2495.3650.1460.448
8demosHist. Gradient Boostinggt_group_ca24145.1496.402-0.1310.340
9demosHist. Gradient Boostinggt_interpersonal_ca24144.8336.086-0.0990.402
22employmentHist. Gradient Boostinggt_group_ca24155.0506.341-0.1090.394
23employmentHist. Gradient Boostinggt_interpersonal_ca24154.8456.115-0.1090.402
36geoHist. Gradient Boostinggt_group_ca24175.0136.302-0.0960.398
37geoHist. Gradient Boostinggt_interpersonal_ca24174.7335.950-0.0500.436
50transitHist. Gradient Boostinggt_group_ca241134.8736.106-0.0290.452
51transitHist. Gradient Boostinggt_interpersonal_ca241134.5185.7670.0130.481
4demosk-NNgt_group_ca24145.8297.219-0.4380.365
5demosk-NNgt_interpersonal_ca24145.2676.782-0.3650.390
18employmentk-NNgt_group_ca24155.8027.308-0.4740.373
19employmentk-NNgt_interpersonal_ca24155.2216.782-0.3650.423
32geok-NNgt_group_ca24175.7417.119-0.3980.378
33geok-NNgt_interpersonal_ca24175.0046.437-0.2290.452
46transitk-NNgt_group_ca241135.2096.435-0.1430.419
47transitk-NNgt_interpersonal_ca241134.7506.157-0.1250.481
12demosNeural net (MLP)gt_group_ca24144.9046.048-0.0090.274
13demosNeural net (MLP)gt_interpersonal_ca24144.5205.7070.0340.369
26employmentNeural net (MLP)gt_group_ca24154.7795.9510.0230.315
27employmentNeural net (MLP)gt_interpersonal_ca24154.4125.6170.0640.440
40geoNeural net (MLP)gt_group_ca24174.8235.9760.0150.332
41geoNeural net (MLP)gt_interpersonal_ca24174.4045.6270.0610.452
54transitNeural net (MLP)gt_group_ca241134.5945.5970.1360.427
55transitNeural net (MLP)gt_interpersonal_ca241134.3115.3760.1420.407
6demosRandom Forestgt_group_ca24145.7177.043-0.3690.365
7demosRandom Forestgt_interpersonal_ca24145.2176.638-0.3070.369
20employmentRandom Forestgt_group_ca24155.5926.958-0.3360.365
21employmentRandom Forestgt_interpersonal_ca24155.1066.488-0.2490.373
34geoRandom Forestgt_group_ca24175.0466.375-0.1210.419
35geoRandom Forestgt_interpersonal_ca24174.4545.856-0.0170.485
48transitRandom Forestgt_group_ca241134.6765.8240.0640.481
49transitRandom Forestgt_interpersonal_ca241134.2755.5060.1010.531
0demosRidgegt_group_ca24144.9666.168-0.0500.328
1demosRidgegt_interpersonal_ca24144.6735.912-0.0370.357
14employmentRidgegt_group_ca24154.6555.9110.0360.452
15employmentRidgegt_interpersonal_ca24154.3875.6380.0570.456
28geoRidgegt_group_ca24174.6785.9020.0390.452
29geoRidgegt_interpersonal_ca24174.3475.5640.0810.440
42transitRidgegt_group_ca241134.4935.6050.1330.481
43transitRidgegt_interpersonal_ca241134.2695.3990.1350.469
10demosXGBoostgt_group_ca24145.8837.476-0.5420.324
11demosXGBoostgt_interpersonal_ca24145.5396.986-0.4480.390
24employmentXGBoostgt_group_ca24155.8227.382-0.5040.382
25employmentXGBoostgt_interpersonal_ca24155.5006.941-0.4290.390
38geoXGBoostgt_group_ca24175.2186.608-0.2050.423
39geoXGBoostgt_interpersonal_ca24174.8766.351-0.1970.444
52transitXGBoostgt_group_ca241134.9306.182-0.0550.465
53transitXGBoostgt_interpersonal_ca241134.5255.816-0.0040.481
\n", "
" ], "text/plain": [ " tier model_label target n_samples \\\n", "2 demos Elastic Net gt_group_ca 241 \n", "3 demos Elastic Net gt_interpersonal_ca 241 \n", "16 employment Elastic Net gt_group_ca 241 \n", "17 employment Elastic Net gt_interpersonal_ca 241 \n", "30 geo Elastic Net gt_group_ca 241 \n", "31 geo Elastic Net gt_interpersonal_ca 241 \n", "44 transit Elastic Net gt_group_ca 241 \n", "45 transit Elastic Net gt_interpersonal_ca 241 \n", "8 demos Hist. Gradient Boosting gt_group_ca 241 \n", "9 demos Hist. Gradient Boosting gt_interpersonal_ca 241 \n", "22 employment Hist. Gradient Boosting gt_group_ca 241 \n", "23 employment Hist. Gradient Boosting gt_interpersonal_ca 241 \n", "36 geo Hist. Gradient Boosting gt_group_ca 241 \n", "37 geo Hist. Gradient Boosting gt_interpersonal_ca 241 \n", "50 transit Hist. Gradient Boosting gt_group_ca 241 \n", "51 transit Hist. Gradient Boosting gt_interpersonal_ca 241 \n", "4 demos k-NN gt_group_ca 241 \n", "5 demos k-NN gt_interpersonal_ca 241 \n", "18 employment k-NN gt_group_ca 241 \n", "19 employment k-NN gt_interpersonal_ca 241 \n", "32 geo k-NN gt_group_ca 241 \n", "33 geo k-NN gt_interpersonal_ca 241 \n", "46 transit k-NN gt_group_ca 241 \n", "47 transit k-NN gt_interpersonal_ca 241 \n", "12 demos Neural net (MLP) gt_group_ca 241 \n", "13 demos Neural net (MLP) gt_interpersonal_ca 241 \n", "26 employment Neural net (MLP) gt_group_ca 241 \n", "27 employment Neural net (MLP) gt_interpersonal_ca 241 \n", "40 geo Neural net (MLP) gt_group_ca 241 \n", "41 geo Neural net (MLP) gt_interpersonal_ca 241 \n", "54 transit Neural net (MLP) gt_group_ca 241 \n", "55 transit Neural net (MLP) gt_interpersonal_ca 241 \n", "6 demos Random Forest gt_group_ca 241 \n", "7 demos Random Forest gt_interpersonal_ca 241 \n", "20 employment Random Forest gt_group_ca 241 \n", "21 employment Random Forest gt_interpersonal_ca 241 \n", "34 geo Random Forest gt_group_ca 241 \n", "35 geo Random Forest gt_interpersonal_ca 241 \n", "48 transit Random Forest gt_group_ca 241 \n", "49 transit Random Forest gt_interpersonal_ca 241 \n", "0 demos Ridge gt_group_ca 241 \n", "1 demos Ridge gt_interpersonal_ca 241 \n", "14 employment Ridge gt_group_ca 241 \n", "15 employment Ridge gt_interpersonal_ca 241 \n", "28 geo Ridge gt_group_ca 241 \n", "29 geo Ridge gt_interpersonal_ca 241 \n", "42 transit Ridge gt_group_ca 241 \n", "43 transit Ridge gt_interpersonal_ca 241 \n", "10 demos XGBoost gt_group_ca 241 \n", "11 demos XGBoost gt_interpersonal_ca 241 \n", "24 employment XGBoost gt_group_ca 241 \n", "25 employment XGBoost gt_interpersonal_ca 241 \n", "38 geo XGBoost gt_group_ca 241 \n", "39 geo XGBoost gt_interpersonal_ca 241 \n", "52 transit XGBoost gt_group_ca 241 \n", "53 transit XGBoost gt_interpersonal_ca 241 \n", "\n", " n_features mae rmse r2 band_acc \n", "2 4 4.936 6.106 -0.029 0.311 \n", "3 4 4.648 5.865 -0.020 0.361 \n", "16 5 4.646 5.869 0.050 0.456 \n", "17 5 4.396 5.615 0.065 0.440 \n", "30 7 4.694 5.888 0.043 0.465 \n", "31 7 4.370 5.564 0.081 0.456 \n", "44 13 4.506 5.574 0.143 0.469 \n", "45 13 4.249 5.365 0.146 0.448 \n", "8 4 5.149 6.402 -0.131 0.340 \n", "9 4 4.833 6.086 -0.099 0.402 \n", "22 5 5.050 6.341 -0.109 0.394 \n", "23 5 4.845 6.115 -0.109 0.402 \n", "36 7 5.013 6.302 -0.096 0.398 \n", "37 7 4.733 5.950 -0.050 0.436 \n", "50 13 4.873 6.106 -0.029 0.452 \n", "51 13 4.518 5.767 0.013 0.481 \n", "4 4 5.829 7.219 -0.438 0.365 \n", "5 4 5.267 6.782 -0.365 0.390 \n", "18 5 5.802 7.308 -0.474 0.373 \n", "19 5 5.221 6.782 -0.365 0.423 \n", "32 7 5.741 7.119 -0.398 0.378 \n", "33 7 5.004 6.437 -0.229 0.452 \n", "46 13 5.209 6.435 -0.143 0.419 \n", "47 13 4.750 6.157 -0.125 0.481 \n", "12 4 4.904 6.048 -0.009 0.274 \n", "13 4 4.520 5.707 0.034 0.369 \n", "26 5 4.779 5.951 0.023 0.315 \n", "27 5 4.412 5.617 0.064 0.440 \n", "40 7 4.823 5.976 0.015 0.332 \n", "41 7 4.404 5.627 0.061 0.452 \n", "54 13 4.594 5.597 0.136 0.427 \n", "55 13 4.311 5.376 0.142 0.407 \n", "6 4 5.717 7.043 -0.369 0.365 \n", "7 4 5.217 6.638 -0.307 0.369 \n", "20 5 5.592 6.958 -0.336 0.365 \n", "21 5 5.106 6.488 -0.249 0.373 \n", "34 7 5.046 6.375 -0.121 0.419 \n", "35 7 4.454 5.856 -0.017 0.485 \n", "48 13 4.676 5.824 0.064 0.481 \n", "49 13 4.275 5.506 0.101 0.531 \n", "0 4 4.966 6.168 -0.050 0.328 \n", "1 4 4.673 5.912 -0.037 0.357 \n", "14 5 4.655 5.911 0.036 0.452 \n", "15 5 4.387 5.638 0.057 0.456 \n", "28 7 4.678 5.902 0.039 0.452 \n", "29 7 4.347 5.564 0.081 0.440 \n", "42 13 4.493 5.605 0.133 0.481 \n", "43 13 4.269 5.399 0.135 0.469 \n", "10 4 5.883 7.476 -0.542 0.324 \n", "11 4 5.539 6.986 -0.448 0.390 \n", "24 5 5.822 7.382 -0.504 0.382 \n", "25 5 5.500 6.941 -0.429 0.390 \n", "38 7 5.218 6.608 -0.205 0.423 \n", "39 7 4.876 6.351 -0.197 0.444 \n", "52 13 4.930 6.182 -0.055 0.465 \n", "53 13 4.525 5.816 -0.004 0.481 " ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "metrics.sort_values([\"model\", \"tier\", \"target\"])[\n", " [\n", " \"tier\",\n", " \"model_label\",\n", " \"target\",\n", " \"n_samples\",\n", " \"n_features\",\n", " \"mae\",\n", " \"rmse\",\n", " \"r2\",\n", " \"band_acc\",\n", " ]\n", "]\n" ] }, { "cell_type": "code", "execution_count": 5, "id": "cell08", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:32.028404Z", "iopub.status.busy": "2026-08-05T22:41:32.028337Z", "iopub.status.idle": "2026-08-05T22:41:32.041372Z", "shell.execute_reply": "2026-08-05T22:41:32.040777Z" } }, "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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
modelmodel_labeltiern_samplesmodel_familymae_grouprmse_groupr2_groupexact_acc_groupband_acc_groupmean_band_distance_groupmean_norm_score_distance_groupmean_norm_band_distance_groupmae_interpersonalrmse_interpersonalr2_interpersonalexact_acc_interpersonalband_acc_interpersonalmean_band_distance_interpersonalmean_norm_score_distance_interpersonalmean_norm_band_distance_interpersonal
0ridgeRidgedemos241linear4.9666.168-0.0500.0460.3280.7260.2070.3634.6735.912-0.0370.0540.3570.6970.1950.349
1ridgeRidgeemployment241linear4.6555.9110.0360.0750.4520.6060.1940.3034.3875.6380.0570.0750.4560.6020.1830.301
2ridgeRidgegeo241linear4.6785.9020.0390.0830.4520.5980.1950.2994.3475.5640.0810.0710.4400.6060.1810.303
3ridgeRidgetransit241linear4.4935.6050.1330.0710.4810.5680.1870.2844.2695.3990.1350.0580.4690.5730.1780.286
4elastic_netElastic Netdemos241linear4.9366.106-0.0290.0410.3110.7220.2060.3614.6485.865-0.0200.0540.3610.6890.1940.344
5elastic_netElastic Netemployment241linear4.6465.8690.0500.0750.4560.6020.1940.3014.3965.6150.0650.0620.4400.6180.1830.309
6elastic_netElastic Netgeo241linear4.6945.8880.0430.0750.4650.5810.1960.2904.3705.5640.0810.0710.4560.5850.1820.293
7elastic_netElastic Nettransit241linear4.5065.5740.1430.0540.4690.5680.1880.2844.2495.3650.1460.0660.4480.5810.1770.290
8knnk-NNdemos241instance5.8297.219-0.4380.0460.3650.7630.2430.3825.2676.782-0.3650.0460.3900.7260.2190.363
9knnk-NNemployment241instance5.8027.308-0.4740.0500.3730.7800.2420.3905.2216.782-0.3650.0500.4230.6930.2180.346
10knnk-NNgeo241instance5.7417.119-0.3980.0500.3780.7800.2390.3905.0046.437-0.2290.0790.4520.6560.2090.328
11knnk-NNtransit241instance5.2096.435-0.1430.0540.4190.6760.2170.3384.7506.157-0.1250.0870.4810.5980.1980.299
12random_forestRandom Forestdemos241tree_ensemble5.7177.043-0.3690.0500.3650.7800.2380.3905.2176.638-0.3070.0910.3690.7430.2170.371
13random_forestRandom Forestemployment241tree_ensemble5.5926.958-0.3360.0370.3650.7590.2330.3805.1066.488-0.2490.0750.3730.7260.2130.363
14random_forestRandom Forestgeo241tree_ensemble5.0466.375-0.1210.0790.4190.6800.2100.3404.4545.856-0.0170.1080.4850.5980.1860.299
15random_forestRandom Foresttransit241tree_ensemble4.6765.8240.0640.0620.4810.5730.1950.2864.2755.5060.1010.0750.5310.5150.1780.257
16hist_gradient_boostingHist. Gradient Boostingdemos241boosting5.1496.402-0.1310.0710.3400.7470.2150.3734.8336.086-0.0990.0660.4020.6560.2010.328
17hist_gradient_boostingHist. Gradient Boostingemployment241boosting5.0506.341-0.1090.0460.3940.6930.2100.3464.8456.115-0.1090.0460.4020.6890.2020.344
18hist_gradient_boostingHist. Gradient Boostinggeo241boosting5.0136.302-0.0960.0710.3980.6890.2090.3444.7335.950-0.0500.0580.4360.6270.1970.313
19hist_gradient_boostingHist. Gradient Boostingtransit241boosting4.8736.106-0.0290.0620.4520.6390.2030.3204.5185.7670.0130.0830.4810.5980.1880.299
20xgboostXGBoostdemos241boosting5.8837.476-0.5420.0830.3240.8380.2450.4195.5396.986-0.4480.0500.3900.7390.2310.369
21xgboostXGBoostemployment241boosting5.8227.382-0.5040.0460.3820.7630.2430.3825.5006.941-0.4290.0620.3900.7390.2290.369
22xgboostXGBoostgeo241boosting5.2186.608-0.2050.0830.4230.6680.2170.3344.8766.351-0.1970.0950.4440.6510.2030.326
23xgboostXGBoosttransit241boosting4.9306.182-0.0550.0540.4650.6220.2050.3114.5255.816-0.0040.0830.4810.5890.1890.295
24mlpNeural net (MLP)demos241neural4.9046.048-0.0090.0460.2740.7470.2040.3734.5205.7070.0340.0410.3690.6680.1880.334
25mlpNeural net (MLP)employment241neural4.7795.9510.0230.0460.3150.7180.1990.3594.4125.6170.0640.0620.4400.5980.1840.299
26mlpNeural net (MLP)geo241neural4.8235.9760.0150.0410.3320.7180.2010.3594.4045.6270.0610.0790.4520.5850.1830.293
27mlpNeural net (MLP)transit241neural4.5945.5970.1360.0500.4270.5980.1910.2994.3115.3760.1420.0410.4070.6220.1800.311
\n", "
" ], "text/plain": [ " model model_label tier n_samples \\\n", "0 ridge Ridge demos 241 \n", "1 ridge Ridge employment 241 \n", "2 ridge Ridge geo 241 \n", "3 ridge Ridge transit 241 \n", "4 elastic_net Elastic Net demos 241 \n", "5 elastic_net Elastic Net employment 241 \n", "6 elastic_net Elastic Net geo 241 \n", "7 elastic_net Elastic Net transit 241 \n", "8 knn k-NN demos 241 \n", "9 knn k-NN employment 241 \n", "10 knn k-NN geo 241 \n", "11 knn k-NN transit 241 \n", "12 random_forest Random Forest demos 241 \n", "13 random_forest Random Forest employment 241 \n", "14 random_forest Random Forest geo 241 \n", "15 random_forest Random Forest transit 241 \n", "16 hist_gradient_boosting Hist. Gradient Boosting demos 241 \n", "17 hist_gradient_boosting Hist. Gradient Boosting employment 241 \n", "18 hist_gradient_boosting Hist. Gradient Boosting geo 241 \n", "19 hist_gradient_boosting Hist. Gradient Boosting transit 241 \n", "20 xgboost XGBoost demos 241 \n", "21 xgboost XGBoost employment 241 \n", "22 xgboost XGBoost geo 241 \n", "23 xgboost XGBoost transit 241 \n", "24 mlp Neural net (MLP) demos 241 \n", "25 mlp Neural net (MLP) employment 241 \n", "26 mlp Neural net (MLP) geo 241 \n", "27 mlp Neural net (MLP) transit 241 \n", "\n", " model_family mae_group rmse_group r2_group exact_acc_group \\\n", "0 linear 4.966 6.168 -0.050 0.046 \n", "1 linear 4.655 5.911 0.036 0.075 \n", "2 linear 4.678 5.902 0.039 0.083 \n", "3 linear 4.493 5.605 0.133 0.071 \n", "4 linear 4.936 6.106 -0.029 0.041 \n", "5 linear 4.646 5.869 0.050 0.075 \n", "6 linear 4.694 5.888 0.043 0.075 \n", "7 linear 4.506 5.574 0.143 0.054 \n", "8 instance 5.829 7.219 -0.438 0.046 \n", "9 instance 5.802 7.308 -0.474 0.050 \n", "10 instance 5.741 7.119 -0.398 0.050 \n", "11 instance 5.209 6.435 -0.143 0.054 \n", "12 tree_ensemble 5.717 7.043 -0.369 0.050 \n", "13 tree_ensemble 5.592 6.958 -0.336 0.037 \n", "14 tree_ensemble 5.046 6.375 -0.121 0.079 \n", "15 tree_ensemble 4.676 5.824 0.064 0.062 \n", "16 boosting 5.149 6.402 -0.131 0.071 \n", "17 boosting 5.050 6.341 -0.109 0.046 \n", "18 boosting 5.013 6.302 -0.096 0.071 \n", "19 boosting 4.873 6.106 -0.029 0.062 \n", "20 boosting 5.883 7.476 -0.542 0.083 \n", "21 boosting 5.822 7.382 -0.504 0.046 \n", "22 boosting 5.218 6.608 -0.205 0.083 \n", "23 boosting 4.930 6.182 -0.055 0.054 \n", "24 neural 4.904 6.048 -0.009 0.046 \n", "25 neural 4.779 5.951 0.023 0.046 \n", "26 neural 4.823 5.976 0.015 0.041 \n", "27 neural 4.594 5.597 0.136 0.050 \n", "\n", " band_acc_group mean_band_distance_group mean_norm_score_distance_group \\\n", "0 0.328 0.726 0.207 \n", "1 0.452 0.606 0.194 \n", "2 0.452 0.598 0.195 \n", "3 0.481 0.568 0.187 \n", "4 0.311 0.722 0.206 \n", "5 0.456 0.602 0.194 \n", "6 0.465 0.581 0.196 \n", "7 0.469 0.568 0.188 \n", "8 0.365 0.763 0.243 \n", "9 0.373 0.780 0.242 \n", "10 0.378 0.780 0.239 \n", "11 0.419 0.676 0.217 \n", "12 0.365 0.780 0.238 \n", "13 0.365 0.759 0.233 \n", "14 0.419 0.680 0.210 \n", "15 0.481 0.573 0.195 \n", "16 0.340 0.747 0.215 \n", "17 0.394 0.693 0.210 \n", "18 0.398 0.689 0.209 \n", "19 0.452 0.639 0.203 \n", "20 0.324 0.838 0.245 \n", "21 0.382 0.763 0.243 \n", "22 0.423 0.668 0.217 \n", "23 0.465 0.622 0.205 \n", "24 0.274 0.747 0.204 \n", "25 0.315 0.718 0.199 \n", "26 0.332 0.718 0.201 \n", "27 0.427 0.598 0.191 \n", "\n", " mean_norm_band_distance_group mae_interpersonal rmse_interpersonal \\\n", "0 0.363 4.673 5.912 \n", "1 0.303 4.387 5.638 \n", "2 0.299 4.347 5.564 \n", "3 0.284 4.269 5.399 \n", "4 0.361 4.648 5.865 \n", "5 0.301 4.396 5.615 \n", "6 0.290 4.370 5.564 \n", "7 0.284 4.249 5.365 \n", "8 0.382 5.267 6.782 \n", "9 0.390 5.221 6.782 \n", "10 0.390 5.004 6.437 \n", "11 0.338 4.750 6.157 \n", "12 0.390 5.217 6.638 \n", "13 0.380 5.106 6.488 \n", "14 0.340 4.454 5.856 \n", "15 0.286 4.275 5.506 \n", "16 0.373 4.833 6.086 \n", "17 0.346 4.845 6.115 \n", "18 0.344 4.733 5.950 \n", "19 0.320 4.518 5.767 \n", "20 0.419 5.539 6.986 \n", "21 0.382 5.500 6.941 \n", "22 0.334 4.876 6.351 \n", "23 0.311 4.525 5.816 \n", "24 0.373 4.520 5.707 \n", "25 0.359 4.412 5.617 \n", "26 0.359 4.404 5.627 \n", "27 0.299 4.311 5.376 \n", "\n", " r2_interpersonal exact_acc_interpersonal band_acc_interpersonal \\\n", "0 -0.037 0.054 0.357 \n", "1 0.057 0.075 0.456 \n", "2 0.081 0.071 0.440 \n", "3 0.135 0.058 0.469 \n", "4 -0.020 0.054 0.361 \n", "5 0.065 0.062 0.440 \n", "6 0.081 0.071 0.456 \n", "7 0.146 0.066 0.448 \n", "8 -0.365 0.046 0.390 \n", "9 -0.365 0.050 0.423 \n", "10 -0.229 0.079 0.452 \n", "11 -0.125 0.087 0.481 \n", "12 -0.307 0.091 0.369 \n", "13 -0.249 0.075 0.373 \n", "14 -0.017 0.108 0.485 \n", "15 0.101 0.075 0.531 \n", "16 -0.099 0.066 0.402 \n", "17 -0.109 0.046 0.402 \n", "18 -0.050 0.058 0.436 \n", "19 0.013 0.083 0.481 \n", "20 -0.448 0.050 0.390 \n", "21 -0.429 0.062 0.390 \n", "22 -0.197 0.095 0.444 \n", "23 -0.004 0.083 0.481 \n", "24 0.034 0.041 0.369 \n", "25 0.064 0.062 0.440 \n", "26 0.061 0.079 0.452 \n", "27 0.142 0.041 0.407 \n", "\n", " mean_band_distance_interpersonal mean_norm_score_distance_interpersonal \\\n", "0 0.697 0.195 \n", "1 0.602 0.183 \n", "2 0.606 0.181 \n", "3 0.573 0.178 \n", "4 0.689 0.194 \n", "5 0.618 0.183 \n", "6 0.585 0.182 \n", "7 0.581 0.177 \n", "8 0.726 0.219 \n", "9 0.693 0.218 \n", "10 0.656 0.209 \n", "11 0.598 0.198 \n", "12 0.743 0.217 \n", "13 0.726 0.213 \n", "14 0.598 0.186 \n", "15 0.515 0.178 \n", "16 0.656 0.201 \n", "17 0.689 0.202 \n", "18 0.627 0.197 \n", "19 0.598 0.188 \n", "20 0.739 0.231 \n", "21 0.739 0.229 \n", "22 0.651 0.203 \n", "23 0.589 0.189 \n", "24 0.668 0.188 \n", "25 0.598 0.184 \n", "26 0.585 0.183 \n", "27 0.622 0.180 \n", "\n", " mean_norm_band_distance_interpersonal \n", "0 0.349 \n", "1 0.301 \n", "2 0.303 \n", "3 0.286 \n", "4 0.344 \n", "5 0.309 \n", "6 0.293 \n", "7 0.290 \n", "8 0.363 \n", "9 0.346 \n", "10 0.328 \n", "11 0.299 \n", "12 0.371 \n", "13 0.363 \n", "14 0.299 \n", "15 0.257 \n", "16 0.328 \n", "17 0.344 \n", "18 0.313 \n", "19 0.299 \n", "20 0.369 \n", "21 0.369 \n", "22 0.326 \n", "23 0.295 \n", "24 0.334 \n", "25 0.299 \n", "26 0.293 \n", "27 0.311 " ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "wide = metrics_wide(metrics)\n", "wide\n" ] }, { "cell_type": "code", "execution_count": 6, "id": "cell09", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:32.042586Z", "iopub.status.busy": "2026-08-05T22:41:32.042499Z", "iopub.status.idle": "2026-08-05T22:41:32.058513Z", "shell.execute_reply": "2026-08-05T22:41:32.058052Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Group-CA MAE pivot\n" ] }, { "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", "
tiermodeldemosemploymentgeotransit
0Ridge4.9664.6554.6784.493
1Elastic Net4.9364.6464.6944.506
2k-NN5.8295.8025.7415.209
3Random Forest5.7175.5925.0464.676
4Hist. Gradient Boosting5.1495.0505.0134.873
5XGBoost5.8835.8225.2184.930
6Neural net (MLP)4.9044.7794.8234.594
\n", "
" ], "text/plain": [ "tier model demos employment geo transit\n", "0 Ridge 4.966 4.655 4.678 4.493\n", "1 Elastic Net 4.936 4.646 4.694 4.506\n", "2 k-NN 5.829 5.802 5.741 5.209\n", "3 Random Forest 5.717 5.592 5.046 4.676\n", "4 Hist. Gradient Boosting 5.149 5.050 5.013 4.873\n", "5 XGBoost 5.883 5.822 5.218 4.930\n", "6 Neural net (MLP) 4.904 4.779 4.823 4.594" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Interpersonal-CA MAE pivot\n" ] }, { "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", "
tiermodeldemosemploymentgeotransit
0Ridge4.6734.3874.3474.269
1Elastic Net4.6484.3964.3704.249
2k-NN5.2675.2215.0044.750
3Random Forest5.2175.1064.4544.275
4Hist. Gradient Boosting4.8334.8454.7334.518
5XGBoost5.5395.5004.8764.525
6Neural net (MLP)4.5204.4124.4044.311
\n", "
" ], "text/plain": [ "tier model demos employment geo transit\n", "0 Ridge 4.673 4.387 4.347 4.269\n", "1 Elastic Net 4.648 4.396 4.370 4.249\n", "2 k-NN 5.267 5.221 5.004 4.750\n", "3 Random Forest 5.217 5.106 4.454 4.275\n", "4 Hist. Gradient Boosting 4.833 4.845 4.733 4.518\n", "5 XGBoost 5.539 5.500 4.876 4.525\n", "6 Neural net (MLP) 4.520 4.412 4.404 4.311" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "Best model per tier × target\n" ] }, { "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", "
tiertargetbest_modelbest_model_labelbest_maebest_rmsebest_r2best_band_accrunner_up_maemae_gap_to_secondn_modelsn_samples
0demosgt_group_camlpNeural net (MLP)4.9046.048-0.0090.2744.9360.0327241
1demosgt_interpersonal_camlpNeural net (MLP)4.5205.7070.0340.3694.6480.1287241
2employmentgt_group_caelastic_netElastic Net4.6465.8690.0500.4564.6550.0107241
3employmentgt_interpersonal_caridgeRidge4.3875.6380.0570.4564.3960.0097241
4geogt_group_caridgeRidge4.6785.9020.0390.4524.6940.0177241
5geogt_interpersonal_caridgeRidge4.3475.5640.0810.4404.3700.0237241
6transitgt_group_caridgeRidge4.4935.6050.1330.4814.5060.0127241
7transitgt_interpersonal_caelastic_netElastic Net4.2495.3650.1460.4484.2690.0207241
\n", "
" ], "text/plain": [ " tier target best_model best_model_label best_mae \\\n", "0 demos gt_group_ca mlp Neural net (MLP) 4.904 \n", "1 demos gt_interpersonal_ca mlp Neural net (MLP) 4.520 \n", "2 employment gt_group_ca elastic_net Elastic Net 4.646 \n", "3 employment gt_interpersonal_ca ridge Ridge 4.387 \n", "4 geo gt_group_ca ridge Ridge 4.678 \n", "5 geo gt_interpersonal_ca ridge Ridge 4.347 \n", "6 transit gt_group_ca ridge Ridge 4.493 \n", "7 transit gt_interpersonal_ca elastic_net Elastic Net 4.249 \n", "\n", " best_rmse best_r2 best_band_acc runner_up_mae mae_gap_to_second \\\n", "0 6.048 -0.009 0.274 4.936 0.032 \n", "1 5.707 0.034 0.369 4.648 0.128 \n", "2 5.869 0.050 0.456 4.655 0.010 \n", "3 5.638 0.057 0.456 4.396 0.009 \n", "4 5.902 0.039 0.452 4.694 0.017 \n", "5 5.564 0.081 0.440 4.370 0.023 \n", "6 5.605 0.133 0.481 4.506 0.012 \n", "7 5.365 0.146 0.448 4.269 0.020 \n", "\n", " n_models n_samples \n", "0 7 241 \n", "1 7 241 \n", "2 7 241 \n", "3 7 241 \n", "4 7 241 \n", "5 7 241 \n", "6 7 241 \n", "7 7 241 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "print(\"Group-CA MAE pivot\")\n", "display(mae_pivot(metrics, target=\"gt_group_ca\"))\n", "print(\"Interpersonal-CA MAE pivot\")\n", "display(mae_pivot(metrics, target=\"gt_interpersonal_ca\"))\n", "print(\"Best model per tier × target\")\n", "display(leaderboard(metrics))\n" ] }, { "cell_type": "markdown", "id": "cell10", "metadata": {}, "source": [ "## 4. Visualize suite MAE by tier\n", "\n", "Line plots compare all seven learners; bar charts highlight the transit-tier ranking (best model in UW red).\n" ] }, { "cell_type": "code", "execution_count": 7, "id": "cell11", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:32.059591Z", "iopub.status.busy": "2026-08-05T22:41:32.059509Z", "iopub.status.idle": "2026-08-05T22:41:32.518038Z", "shell.execute_reply": "2026-08-05T22:41:32.517557Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from ca_personas.apa_plotting import apa_axes, apply_apa_style\n", "\n", "apply_apa_style()\n", "tiers = list(RESEARCH_TIERS)\n", "models = list(DEFAULT_MODEL_SUITE)\n", "colors = {\n", " \"ridge\": \"#222222\",\n", " \"elastic_net\": \"#555555\",\n", " \"knn\": \"#888888\",\n", " \"random_forest\": \"#C5050C\",\n", " \"hist_gradient_boosting\": \"#444444\",\n", " \"xgboost\": \"#666666\",\n", " \"mlp\": \"#000000\",\n", "}\n", "markers = {\n", " \"ridge\": \"o\",\n", " \"elastic_net\": \"s\",\n", " \"knn\": \"D\",\n", " \"random_forest\": \"^\",\n", " \"hist_gradient_boosting\": \"v\",\n", " \"xgboost\": \"P\",\n", " \"mlp\": \"X\",\n", "}\n", "\n", "for target, fname, heading in (\n", " (\"gt_group_ca\", \"ml_baseline_mae_group.png\", \"Group CA\"),\n", " (\"gt_interpersonal_ca\", \"ml_baseline_mae_interpersonal.png\", \"Interpersonal CA\"),\n", "):\n", " fig, ax = plt.subplots(figsize=(7.2, 4.2))\n", " for model in models:\n", " sub = metrics[(metrics[\"target\"] == target) & (metrics[\"model\"] == model)].copy()\n", " sub[\"tier\"] = pd.Categorical(sub[\"tier\"], categories=tiers, ordered=True)\n", " sub = sub.sort_values(\"tier\")\n", " ax.plot(\n", " range(len(tiers)),\n", " sub[\"mae\"].tolist(),\n", " marker=markers[model],\n", " color=colors[model],\n", " linewidth=1.8 if model == \"random_forest\" else 1.25,\n", " label=MODEL_LABELS[model],\n", " )\n", " ax.set_xticks(range(len(tiers)))\n", " ax.set_xticklabels(tiers)\n", " ax.set_ylabel(\"Mean absolute error\")\n", " ax.set_xlabel(\"Persona information tier\")\n", " ax.set_ylim(3.8, 6.4)\n", " apa_axes(ax)\n", " ax.legend(frameon=False, fontsize=8, ncol=2, loc=\"upper right\")\n", " ax.text(0.02, 0.98, heading, transform=ax.transAxes, va=\"top\", fontweight=\"bold\")\n", " fig.tight_layout()\n", " fig.savefig(OUT_DIR / fname, dpi=300, bbox_inches=\"tight\", facecolor=\"white\")\n", " fig.savefig(FIG_DIR / fname, dpi=300, bbox_inches=\"tight\", facecolor=\"white\")\n", " plt.show()\n" ] }, { "cell_type": "code", "execution_count": 8, "id": "cell12", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:32.519330Z", "iopub.status.busy": "2026-08-05T22:41:32.519231Z", "iopub.status.idle": "2026-08-05T22:41:32.767567Z", "shell.execute_reply": "2026-08-05T22:41:32.767172Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import numpy as np\n", "\n", "fig, axes = plt.subplots(1, 2, figsize=(10, 4.2), sharey=True)\n", "for ax, target, heading in zip(\n", " axes,\n", " [\"gt_group_ca\", \"gt_interpersonal_ca\"],\n", " [\"Group CA @ transit\", \"Interpersonal CA @ transit\"],\n", "):\n", " sub = metrics[(metrics[\"tier\"] == \"transit\") & (metrics[\"target\"] == target)].copy()\n", " sub[\"model\"] = pd.Categorical(sub[\"model\"], categories=models, ordered=True)\n", " sub = sub.sort_values(\"model\")\n", " best = sub.loc[sub[\"mae\"].idxmin(), \"model\"]\n", " x = np.arange(len(sub))\n", " ax.bar(\n", " x,\n", " sub[\"mae\"],\n", " color=[\"#C5050C\" if m == best else \"#777777\" for m in sub[\"model\"]],\n", " edgecolor=\"black\",\n", " linewidth=0.4,\n", " )\n", " ax.set_xticks(x)\n", " ax.set_xticklabels([MODEL_LABELS[m] for m in sub[\"model\"]], rotation=35, ha=\"right\", fontsize=8)\n", " ax.set_ylabel(\"MAE\" if ax is axes[0] else \"\")\n", " ax.set_xlabel(\"Model\")\n", " ax.set_ylim(0, 6.2)\n", " apa_axes(ax)\n", " ax.text(0.02, 0.98, heading, transform=ax.transAxes, va=\"top\", fontweight=\"bold\")\n", "fig.tight_layout()\n", "fig.savefig(OUT_DIR / \"ml_baseline_mae_transit_bars.png\", dpi=300, bbox_inches=\"tight\", facecolor=\"white\")\n", "fig.savefig(FIG_DIR / \"ml_baseline_mae_transit_bars.png\", dpi=300, bbox_inches=\"tight\", facecolor=\"white\")\n", "plt.show()\n" ] }, { "cell_type": "markdown", "id": "cell13", "metadata": {}, "source": [ "## 5. Per-participant predictions\n" ] }, { "cell_type": "code", "execution_count": 9, "id": "cell14", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:32.770568Z", "iopub.status.busy": "2026-08-05T22:41:32.770218Z", "iopub.status.idle": "2026-08-05T22:41:32.783601Z", "shell.execute_reply": "2026-08-05T22:41:32.783232Z" } }, "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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
participant_idtiermodelmodel_labelmodel_familytargetsidey_truey_prederrorabs_errorscore_distancenorm_score_distancegt_bandpred_bandexact_matchband_matchband_distancenorm_band_distance
59801ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demoselastic_netElastic Netlineargt_group_cagroup7.00015.3918.3918.3918.3910.350lowmoderateFalseFalse10.500
83901ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demoselastic_netElastic Netlineargt_interpersonal_cainterpersonal6.00014.6048.6048.6048.6040.359lowmoderateFalseFalse10.500
204401ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demoshist_gradient_boostingHist. Gradient Boostingboostinggt_group_cagroup7.0009.0962.0962.0962.0960.087lowlowFalseTrue00.000
228501ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demoshist_gradient_boostingHist. Gradient Boostingboostinggt_interpersonal_cainterpersonal6.0008.6342.6342.6342.6340.110lowlowFalseTrue00.000
108001ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosknnk-NNinstancegt_group_cagroup7.0009.2502.2502.2502.2500.094lowlowFalseTrue00.000
132101ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosknnk-NNinstancegt_interpersonal_cainterpersonal6.0009.5003.5003.5003.5000.146lowlowFalseTrue00.000
300801ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosmlpNeural net (MLP)neuralgt_group_cagroup7.00015.1218.1218.1218.1210.338lowmoderateFalseFalse10.500
324901ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosmlpNeural net (MLP)neuralgt_interpersonal_cainterpersonal6.00014.9078.9078.9078.9070.371lowmoderateFalseFalse10.500
156201ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosrandom_forestRandom Foresttree_ensemblegt_group_cagroup7.00010.6213.6213.6213.6210.151lowlowFalseTrue00.000
180301ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosrandom_forestRandom Foresttree_ensemblegt_interpersonal_cainterpersonal6.00011.8175.8175.8175.8170.242lowlowFalseTrue00.000
11601ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosridgeRidgelineargt_group_cagroup7.00015.4448.4448.4448.4440.352lowmoderateFalseFalse10.500
35701ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosridgeRidgelineargt_interpersonal_cainterpersonal6.00014.6308.6308.6308.6300.360lowmoderateFalseFalse10.500
252601ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosxgboostXGBoostboostinggt_group_cagroup7.0008.8641.8641.8641.8640.078lowlowFalseTrue00.000
276701ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...demosxgboostXGBoostboostinggt_interpersonal_cainterpersonal6.0009.3963.3963.3963.3960.142lowlowFalseTrue00.000
397201ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmentelastic_netElastic Netlineargt_group_cagroup7.00014.6447.6447.6447.6440.318lowmoderateFalseFalse10.500
421301ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmentelastic_netElastic Netlineargt_interpersonal_cainterpersonal6.00013.8327.8327.8327.8320.326lowmoderateFalseFalse10.500
541801ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmenthist_gradient_boostingHist. Gradient Boostingboostinggt_group_cagroup7.0009.3422.3422.3422.3420.098lowlowFalseTrue00.000
565901ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmenthist_gradient_boostingHist. Gradient Boostingboostinggt_interpersonal_cainterpersonal6.00010.3034.3034.3034.3030.179lowlowFalseTrue00.000
445401ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmentknnk-NNinstancegt_group_cagroup7.0008.5451.5451.5451.5450.064lowlowFalseTrue00.000
469501ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmentknnk-NNinstancegt_interpersonal_cainterpersonal6.0008.1822.1822.1822.1820.091lowlowFalseTrue00.000
638201ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmentmlpNeural net (MLP)neuralgt_group_cagroup7.00015.4918.4918.4918.4910.354lowmoderateFalseFalse10.500
662301ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmentmlpNeural net (MLP)neuralgt_interpersonal_cainterpersonal6.00015.2049.2049.2049.2040.383lowmoderateFalseFalse10.500
493601ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmentrandom_forestRandom Foresttree_ensemblegt_group_cagroup7.00011.3294.3294.3294.3290.180lowlowFalseTrue00.000
517701ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc...employmentrandom_forestRandom Foresttree_ensemblegt_interpersonal_cainterpersonal6.00011.8615.8615.8615.8610.244lowlowFalseTrue00.000
\n", "
" ], "text/plain": [ " participant_id tier \\\n", "598 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "839 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "2044 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "2285 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "1080 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "1321 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "3008 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "3249 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "1562 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "1803 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "116 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "357 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "2526 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "2767 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... demos \n", "3972 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "4213 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "5418 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "5659 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "4454 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "4695 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "6382 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "6623 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "4936 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "5177 01ed97565030e4d13eba2ecba6f408ba9b1d77f47fa2cc... employment \n", "\n", " model model_label model_family \\\n", "598 elastic_net Elastic Net linear \n", "839 elastic_net Elastic Net linear \n", "2044 hist_gradient_boosting Hist. Gradient Boosting boosting \n", "2285 hist_gradient_boosting Hist. Gradient Boosting boosting \n", "1080 knn k-NN instance \n", "1321 knn k-NN instance \n", "3008 mlp Neural net (MLP) neural \n", "3249 mlp Neural net (MLP) neural \n", "1562 random_forest Random Forest tree_ensemble \n", "1803 random_forest Random Forest tree_ensemble \n", "116 ridge Ridge linear \n", "357 ridge Ridge linear \n", "2526 xgboost XGBoost boosting \n", "2767 xgboost XGBoost boosting \n", "3972 elastic_net Elastic Net linear \n", "4213 elastic_net Elastic Net linear \n", "5418 hist_gradient_boosting Hist. Gradient Boosting boosting \n", "5659 hist_gradient_boosting Hist. Gradient Boosting boosting \n", "4454 knn k-NN instance \n", "4695 knn k-NN instance \n", "6382 mlp Neural net (MLP) neural \n", "6623 mlp Neural net (MLP) neural \n", "4936 random_forest Random Forest tree_ensemble \n", "5177 random_forest Random Forest tree_ensemble \n", "\n", " target side y_true y_pred error abs_error \\\n", "598 gt_group_ca group 7.000 15.391 8.391 8.391 \n", "839 gt_interpersonal_ca interpersonal 6.000 14.604 8.604 8.604 \n", "2044 gt_group_ca group 7.000 9.096 2.096 2.096 \n", "2285 gt_interpersonal_ca interpersonal 6.000 8.634 2.634 2.634 \n", "1080 gt_group_ca group 7.000 9.250 2.250 2.250 \n", "1321 gt_interpersonal_ca interpersonal 6.000 9.500 3.500 3.500 \n", "3008 gt_group_ca group 7.000 15.121 8.121 8.121 \n", "3249 gt_interpersonal_ca interpersonal 6.000 14.907 8.907 8.907 \n", "1562 gt_group_ca group 7.000 10.621 3.621 3.621 \n", "1803 gt_interpersonal_ca interpersonal 6.000 11.817 5.817 5.817 \n", "116 gt_group_ca group 7.000 15.444 8.444 8.444 \n", "357 gt_interpersonal_ca interpersonal 6.000 14.630 8.630 8.630 \n", "2526 gt_group_ca group 7.000 8.864 1.864 1.864 \n", "2767 gt_interpersonal_ca interpersonal 6.000 9.396 3.396 3.396 \n", "3972 gt_group_ca group 7.000 14.644 7.644 7.644 \n", "4213 gt_interpersonal_ca interpersonal 6.000 13.832 7.832 7.832 \n", "5418 gt_group_ca group 7.000 9.342 2.342 2.342 \n", "5659 gt_interpersonal_ca interpersonal 6.000 10.303 4.303 4.303 \n", "4454 gt_group_ca group 7.000 8.545 1.545 1.545 \n", "4695 gt_interpersonal_ca interpersonal 6.000 8.182 2.182 2.182 \n", "6382 gt_group_ca group 7.000 15.491 8.491 8.491 \n", "6623 gt_interpersonal_ca interpersonal 6.000 15.204 9.204 9.204 \n", "4936 gt_group_ca group 7.000 11.329 4.329 4.329 \n", "5177 gt_interpersonal_ca interpersonal 6.000 11.861 5.861 5.861 \n", "\n", " score_distance norm_score_distance gt_band pred_band exact_match \\\n", "598 8.391 0.350 low moderate False \n", "839 8.604 0.359 low moderate False \n", "2044 2.096 0.087 low low False \n", "2285 2.634 0.110 low low False \n", "1080 2.250 0.094 low low False \n", "1321 3.500 0.146 low low False \n", "3008 8.121 0.338 low moderate False \n", "3249 8.907 0.371 low moderate False \n", "1562 3.621 0.151 low low False \n", "1803 5.817 0.242 low low False \n", "116 8.444 0.352 low moderate False \n", "357 8.630 0.360 low moderate False \n", "2526 1.864 0.078 low low False \n", "2767 3.396 0.142 low low False \n", "3972 7.644 0.318 low moderate False \n", "4213 7.832 0.326 low moderate False \n", "5418 2.342 0.098 low low False \n", "5659 4.303 0.179 low low False \n", "4454 1.545 0.064 low low False \n", "4695 2.182 0.091 low low False \n", "6382 8.491 0.354 low moderate False \n", "6623 9.204 0.383 low moderate False \n", "4936 4.329 0.180 low low False \n", "5177 5.861 0.244 low low False \n", "\n", " band_match band_distance norm_band_distance \n", "598 False 1 0.500 \n", "839 False 1 0.500 \n", "2044 True 0 0.000 \n", "2285 True 0 0.000 \n", "1080 True 0 0.000 \n", "1321 True 0 0.000 \n", "3008 False 1 0.500 \n", "3249 False 1 0.500 \n", "1562 True 0 0.000 \n", "1803 True 0 0.000 \n", "116 False 1 0.500 \n", "357 False 1 0.500 \n", "2526 True 0 0.000 \n", "2767 True 0 0.000 \n", "3972 False 1 0.500 \n", "4213 False 1 0.500 \n", "5418 True 0 0.000 \n", "5659 True 0 0.000 \n", "4454 True 0 0.000 \n", "4695 True 0 0.000 \n", "6382 False 1 0.500 \n", "6623 False 1 0.500 \n", "4936 True 0 0.000 \n", "5177 True 0 0.000 " ] }, "execution_count": 9, "metadata": {}, "output_type": "execute_result" } ], "source": [ "predictions.sort_values([\"participant_id\", \"tier\", \"model\", \"side\"]).head(24)\n" ] }, { "cell_type": "markdown", "id": "cell15", "metadata": {}, "source": [ "## 6. Persist artifacts\n" ] }, { "cell_type": "code", "execution_count": 10, "id": "cell16", "metadata": { "execution": { "iopub.execute_input": "2026-08-05T22:41:32.785059Z", "iopub.status.busy": "2026-08-05T22:41:32.784963Z", "iopub.status.idle": "2026-08-05T22:41:32.892082Z", "shell.execute_reply": "2026-08-05T22:41:32.891649Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "predictions → /Users/morningstar/Desktop/Cold_Storage/psych755-jjb/outputs/ml_baseline/ml_baseline_predictions.csv\n", "metrics → /Users/morningstar/Desktop/Cold_Storage/psych755-jjb/outputs/ml_baseline/ml_baseline_metrics.csv\n", "metrics_wide → /Users/morningstar/Desktop/Cold_Storage/psych755-jjb/outputs/ml_baseline/ml_baseline_metrics_wide.csv\n", "leaderboard → /Users/morningstar/Desktop/Cold_Storage/psych755-jjb/outputs/ml_baseline/ml_baseline_leaderboard.csv\n", "mae_pivot_group → /Users/morningstar/Desktop/Cold_Storage/psych755-jjb/outputs/ml_baseline/ml_baseline_mae_pivot_group.csv\n", "mae_pivot_interpersonal → /Users/morningstar/Desktop/Cold_Storage/psych755-jjb/outputs/ml_baseline/ml_baseline_mae_pivot_interpersonal.csv\n", "\n", "Transit-tier winners:\n", " tier target best_model best_model_label best_mae best_rmse best_r2 best_band_acc runner_up_mae mae_gap_to_second n_models n_samples\n", "transit gt_group_ca ridge Ridge 4.493 5.605 0.133 0.481 4.506 0.012 7 241\n", "transit gt_interpersonal_ca elastic_net Elastic Net 4.249 5.365 0.146 0.448 4.269 0.020 7 241\n" ] } ], "source": [ "paths = save_baseline_artifacts(predictions, metrics, OUT_DIR)\n", "for key, path in paths.items():\n", " print(f\"{key:20s} → {path}\")\n", "board = leaderboard(metrics)\n", "print(\"\\nTransit-tier winners:\")\n", "print(board[board[\"tier\"] == \"transit\"].to_string(index=False))\n" ] }, { "cell_type": "markdown", "id": "cell17", "metadata": {}, "source": [ "## 7. How to read these baselines against the LLM\n", "\n", "1. **Use the suite best, not RF alone.** On the full cohort, Ridge / Elastic Net / MLP often beat Random Forest. Compare LLM MAE to the leaderboard winner at each tier (see `docs/ml_baselines.md`).\n", "2. **Same features, same targets.** An LLM persona tier is competitive only if its MAE approaches the best tabular MAE on that tier.\n", "3. **Band accuracy is a second lens.** Even the best models stay near ~0.45–0.48 band accuracy at transit — demographics + mobility do not pin PRCA bands tightly.\n", "4. **Next step.** Run `ca-personas compare` / `notebooks/ml_vs_llm_comparison.ipynb` to place LLM agents on the same metric table.\n" ] } ], "metadata": { "kernelspec": { "display_name": "Python (.venv psych755-jjb)", "language": "python", "name": "psych755-jjb" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.15" } }, "nbformat": 4, "nbformat_minor": 5 }