ResearchProject/FESurrogateModelTutorial/notebooks/05_mlp_surrogate.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "69c9cc92",
   "metadata": {},
   "source": [
    "# MLP Neural Network Surrogate\n",
    "\n",
    "MLP? scaled input? target? ??? ???? ??? ??? ????.\n",
    "\n",
    "?? ?? notebook? ?? dataset, target, train/test split seed? ????."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "eed089f5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import json\n",
    "import warnings\n",
    "from pathlib import Path\n",
    "\n",
    "import pandas as pd\n",
    "from sklearn.exceptions import ConvergenceWarning\n",
    "\n",
    "from femsurrogate.data.schema import DEFAULT_RANDOM_SEED, PARAMETER_COLUMNS\n",
    "from femsurrogate.plotting.diagnostics import plot_parity, plot_residuals\n",
    "from femsurrogate.surrogates.common import evaluate_model, metrics_to_dict, split_dataset\n",
    "from femsurrogate.surrogates.registry import make_model\n",
    "\n",
    "ROOT = Path.cwd().resolve()\n",
    "if not (ROOT / \"pyproject.toml\").exists():\n",
    "    ROOT = ROOT.parent\n",
    "assert (ROOT / \"pyproject.toml\").exists(), ROOT\n",
    "DATASET_PATH = ROOT / \"data\" / \"reference\" / \"beam2d_lhs_300.csv\"\n",
    "RESULTS_DIR = ROOT / \"reports\" / \"results\"\n",
    "PREDICTIONS_DIR = ROOT / \"reports\" / \"predictions\"\n",
    "FIGURES_DIR = ROOT / \"reports\" / \"figures\"\n",
    "for directory in [RESULTS_DIR, PREDICTIONS_DIR, FIGURES_DIR]:\n",
    "    directory.mkdir(parents=True, exist_ok=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b2547495",
   "metadata": {},
   "source": [
    "## Dataset? split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "039279d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = pd.read_csv(DATASET_PATH)\n",
    "target_column = \"tip_uy_m\"\n",
    "split = split_dataset(\n",
    "    dataset,\n",
    "    feature_columns=list(PARAMETER_COLUMNS),\n",
    "    target_column=target_column,\n",
    "    test_size=0.2,\n",
    "    random_state=DEFAULT_RANDOM_SEED,\n",
    ")\n",
    "\n",
    "len(split.X_train), len(split.X_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "71b3323f",
   "metadata": {},
   "source": [
    "## ?? ??? ??"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "605253ef",
   "metadata": {},
   "outputs": [],
   "source": [
    "MODEL_NAME = \"mlp\"\n",
    "model = make_model(MODEL_NAME, random_state=DEFAULT_RANDOM_SEED, **{'max_iter': 500})\n",
    "\n",
    "with warnings.catch_warnings():\n",
    "    warnings.filterwarnings(\"ignore\", category=ConvergenceWarning)\n",
    "    result = evaluate_model(\n",
    "        model,\n",
    "        split.X_train,\n",
    "        split.X_test,\n",
    "        split.y_train,\n",
    "        split.y_test,\n",
    "        model_name=MODEL_NAME,\n",
    "        target_column=target_column,\n",
    "    )\n",
    "\n",
    "metrics = metrics_to_dict(result.metrics)\n",
    "metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f70ad9fd",
   "metadata": {},
   "source": [
    "## ?? ??"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fb08797b",
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics_path = RESULTS_DIR / f\"{MODEL_NAME}_metrics.json\"\n",
    "predictions_path = PREDICTIONS_DIR / f\"{MODEL_NAME}_predictions.csv\"\n",
    "\n",
    "metrics_path.write_text(json.dumps(metrics, indent=2), encoding=\"utf-8\")\n",
    "result.predictions.to_csv(predictions_path, index=False)\n",
    "\n",
    "{\"metrics_path\": str(metrics_path), \"predictions_path\": str(predictions_path)}"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "39a34166",
   "metadata": {},
   "source": [
    "## ?? plot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "694a1081",
   "metadata": {},
   "outputs": [],
   "source": [
    "parity_fig = plot_parity(result.predictions, title=f\"{MODEL_NAME} parity\")\n",
    "residual_fig = plot_residuals(result.predictions, title=f\"{MODEL_NAME} residuals\")\n",
    "parity_fig.savefig(FIGURES_DIR / f\"{MODEL_NAME}_parity.png\", dpi=150, bbox_inches=\"tight\")\n",
    "residual_fig.savefig(FIGURES_DIR / f\"{MODEL_NAME}_residuals.png\", dpi=150, bbox_inches=\"tight\")\n",
    "parity_fig"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "name": "python",
   "pygments_lexer": "ipython3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}