{ "cells": [ { "cell_type": "code", "execution_count": 1, "id": "08e53f3f", "metadata": {}, "outputs": [ { "name": "stderr", "output_type": "stream", "text": [ "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n", "I0000 00:00:1788414583.218315 23312 port.cc:153] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n", "I0000 00:00:1788414583.219956 23312 cudart_stub.cc:31] Could not find cuda drivers on your machine, GPU will not be used.\n", "I0000 00:00:1788414583.489015 23312 cpu_feature_guard.cc:227] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n", "To enable the following instructions: AVX2 AVX_VNNI AVX_VNNI_INT8 AVX_NE_CONVERT FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n", "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n", "I0000 00:00:1788414585.253354 23312 port.cc:153] oneDNN custom operations are on. You may see slightly different numerical results due to floating-point round-off errors from different computation orders. To turn them off, set the environment variable `TF_ENABLE_ONEDNN_OPTS=0`.\n", "I0000 00:00:1788414585.253946 23312 cudart_stub.cc:31] Could not find cuda drivers on your machine, GPU will not be used.\n" ] } ], "source": [ "#svm\n", "from tensorflow.keras import datasets\n", "from sklearn.model_selection import train_test_split\n", "import numpy as np\n", "import cv2" ] }, { "cell_type": "code", "execution_count": 2, "id": "b7c72748", "metadata": {}, "outputs": [], "source": [ "TRAIN_SIZE = 5000\n", "TEST_SIZE = 1000\n", "\n", "(x_train_raw, y_train_raw), (x_test_raw, y_test_raw) = datasets.mnist.load_data()\n", "\n", "x_train_sub, _, y_train_sub, _ = train_test_split(\n", " x_train_raw, y_train_raw,\n", " train_size=TRAIN_SIZE,\n", " stratify=y_train_raw,\n", " random_state=42\n", ")\n", "\n", "x_test_sub, _, y_test_sub, _ = train_test_split(\n", " x_test_raw, y_test_raw,\n", " train_size=TEST_SIZE,\n", " stratify=y_test_raw,\n", " random_state=42\n", ")\n", "\n", "x_train_eq = np.array([cv2.equalizeHist(img) for img in x_train_sub])\n", "x_test_eq = np.array([cv2.equalizeHist(img) for img in x_test_sub])\n", "\n", "\n", "x_train_flat = x_train_eq.reshape(-1, 28 * 28) / 255.0\n", "x_test_flat = x_test_eq.reshape(-1, 28 * 28) / 255.0\n" ] }, { "cell_type": "code", "execution_count": 3, "id": "b944ab65", "metadata": {}, "outputs": [ { "data": { "text/html": [ "
GridSearchCV(cv=5, estimator=SVC(kernel='linear'), n_jobs=-1,\n",
       "             param_grid={'C': [0.1, 1, 10, 100]}, scoring='accuracy')
In a Jupyter environment, please rerun this cell to show the HTML representation or trust the notebook.
On GitHub, the HTML representation is unable to render, please try loading this page with nbviewer.org.
" ], "text/plain": [ "GridSearchCV(cv=5, estimator=SVC(kernel='linear'), n_jobs=-1,\n", " param_grid={'C': [0.1, 1, 10, 100]}, scoring='accuracy')" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from sklearn.model_selection import train_test_split, GridSearchCV, cross_validate\n", "from sklearn.svm import SVC\n", "\n", "svc = SVC(kernel=\"linear\")\n", "grid_search = GridSearchCV(svc, {\"C\": [0.1, 1, 10, 100]}, cv=5, scoring=\"accuracy\", n_jobs=-1)\n", "grid_search.fit(x_train_flat, y_train_sub)" ] }, { "cell_type": "code", "execution_count": 4, "id": "62876e40", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Kernel Best C Best Gamma Best Degree Best Acc\n", "Linear 0.1 N/A N/A 0.9132\n" ] } ], "source": [ "import pandas as pd\n", "\n", "best_estimators = {}\n", "\n", "best_model = grid_search.best_estimator_\n", "best_params = grid_search.best_params_\n", "best_score = grid_search.best_score_\n", "\n", "best_estimators[\"Linear\"] = best_model\n", "\n", "table1_data = []\n", "table1_data.append({\n", " \"Kernel\" : \"Linear\",\n", " \"Best C\": best_params.get(\"C\", \"N/A\"),\n", " \"Best Gamma\": best_params.get(\"gamma\", \"N/A\"),\n", " \"Best Degree\": best_params.get(\"degree\", \"N/A\"),\n", " \"Best Acc\": best_score\n", "})\n", "\n", "df_table1 = pd.DataFrame(table1_data)\n", "print(df_table1.to_string(index=False))" ] }, { "cell_type": "code", "execution_count": 5, "id": "a25eb0ff", "metadata": {}, "outputs": [], "source": [ "table2_data = []\n", "\n", "cv_res = cross_validate(best_model, x_train_flat, y_train_sub, \n", " cv=5, scoring=[\"accuracy\", \"recall_weighted\", \"f1_weighted\"])" ] }, { "cell_type": "code", "execution_count": 6, "id": "8d088ab3", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "{'fit_time': array([1.05809426, 1.02123737, 1.02057838, 1.03010726, 0.9899869 ]),\n", " 'score_time': array([0.40557742, 0.40280175, 0.40119481, 0.40215659, 0.39224553]),\n", " 'test_accuracy': array([0.907, 0.916, 0.924, 0.917, 0.902]),\n", " 'test_recall_weighted': array([0.907, 0.916, 0.924, 0.917, 0.902]),\n", " 'test_f1_weighted': array([0.90672196, 0.91580509, 0.92439618, 0.91673165, 0.90117893])}" ] }, "execution_count": 6, "metadata": {}, "output_type": "execute_result" } ], "source": [ "cv_res" ] } ], "metadata": { "kernelspec": { "display_name": ".venv", "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": "3.13.12" } }, "nbformat": 4, "nbformat_minor": 5 }