{ "cells": [ { "cell_type": "code", "execution_count": 21, "id": "8eb822ac-ff27-4475-9aa6-374e2b7e77e6", "metadata": {}, "outputs": [], "source": [ "import os, json\n", "import numpy as np\n", "import pandas as pd\n", "\n", "from sklearn.model_selection import train_test_split, StratifiedKFold, GridSearchCV\n", "from sklearn.preprocessing import StandardScaler, OneHotEncoder\n", "from sklearn.compose import ColumnTransformer\n", "from sklearn.pipeline import Pipeline\n", "from sklearn.impute import SimpleImputer\n", "from sklearn.metrics import accuracy_score, f1_score, roc_auc_score, classification_report, confusion_matrix\n", "\n", "from sklearn.linear_model import LogisticRegression\n", "from sklearn.svm import SVC\n", "from sklearn.ensemble import RandomForestClassifier\n", "from sklearn.neighbors import KNeighborsClassifier\n", "\n", "from sklearn.feature_selection import SelectKBest, f_classif, RFE\n", "\n", "RANDOM_STATE = 42\n", "CV_SPLITS = 5\n", "FULL_NAME = \"Atheer_Mohammed_AlShahrani\"\n", "\n", "RAW_CSV_PATH = \"Breast_cancer_dataset.csv\" \n", "CLEAN_CSV_PATH = f\"Breast_cancer_clean_{FULL_NAME}.csv\"\n", "\n", "REPORTS_DIR = f\"final_project_outputs_{FULL_NAME}\"\n", "os.makedirs(REPORTS_DIR, exist_ok=True)\n" ] }, { "cell_type": "code", "execution_count": 22, "id": "9f4fd296-c1b8-4692-881c-8bc8f16e364a", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Target column: diagnosis\n", "Numeric features: 31 | Categorical features: 0\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", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
diagnosisradius_meantexture_meanperimeter_meanarea_meansmoothness_meancompactness_meanconcavity_meanconcave_points_meansymmetry_mean...radius_worsttexture_worstperimeter_worstarea_worstsmoothness_worstcompactness_worstconcavity_worstconcave_points_worstsymmetry_worstfractal_dimension_worst
0117.9910.38122.801001.00.118400.277600.30010.147100.2419...25.3817.33184.602019.00.16220.66560.71190.26540.46010.11890
1120.5717.77132.901326.00.084740.078640.08690.070170.1812...24.9923.41158.801956.00.12380.18660.24160.18600.27500.08902
2119.6921.25130.001203.00.109600.159900.19740.127900.2069...23.5725.53152.501709.00.14440.42450.45040.24300.36130.08758
3111.4220.3877.58386.10.142500.283900.24140.105200.2597...14.9126.5098.87567.70.20980.86630.68690.25750.66380.17300
4120.2914.34135.101297.00.100300.132800.19800.104300.1809...22.5416.67152.201575.00.13740.20500.40000.16250.23640.07678
\n", "

5 rows × 31 columns

\n", "
" ], "text/plain": [ " diagnosis radius_mean texture_mean perimeter_mean area_mean \\\n", "0 1 17.99 10.38 122.80 1001.0 \n", "1 1 20.57 17.77 132.90 1326.0 \n", "2 1 19.69 21.25 130.00 1203.0 \n", "3 1 11.42 20.38 77.58 386.1 \n", "4 1 20.29 14.34 135.10 1297.0 \n", "\n", " smoothness_mean compactness_mean concavity_mean concave_points_mean \\\n", "0 0.11840 0.27760 0.3001 0.14710 \n", "1 0.08474 0.07864 0.0869 0.07017 \n", "2 0.10960 0.15990 0.1974 0.12790 \n", "3 0.14250 0.28390 0.2414 0.10520 \n", "4 0.10030 0.13280 0.1980 0.10430 \n", "\n", " symmetry_mean ... radius_worst texture_worst perimeter_worst \\\n", "0 0.2419 ... 25.38 17.33 184.60 \n", "1 0.1812 ... 24.99 23.41 158.80 \n", "2 0.2069 ... 23.57 25.53 152.50 \n", "3 0.2597 ... 14.91 26.50 98.87 \n", "4 0.1809 ... 22.54 16.67 152.20 \n", "\n", " area_worst smoothness_worst compactness_worst concavity_worst \\\n", "0 2019.0 0.1622 0.6656 0.7119 \n", "1 1956.0 0.1238 0.1866 0.2416 \n", "2 1709.0 0.1444 0.4245 0.4504 \n", "3 567.7 0.2098 0.8663 0.6869 \n", "4 1575.0 0.1374 0.2050 0.4000 \n", "\n", " concave_points_worst symmetry_worst fractal_dimension_worst \n", "0 0.2654 0.4601 0.11890 \n", "1 0.1860 0.2750 0.08902 \n", "2 0.2430 0.3613 0.08758 \n", "3 0.2575 0.6638 0.17300 \n", "4 0.1625 0.2364 0.07678 \n", "\n", "[5 rows x 31 columns]" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "def normalize_col(c: str) -> str:\n", " return (\n", " str(c).strip()\n", " .replace(\"\\n\", \" \").replace(\"\\r\", \" \")\n", " .replace(\"/\", \"_\").replace(\"\\\\\", \"_\")\n", " .replace(\" \", \"_\").lower()\n", " )\n", "\n", "def load_and_clean(path: str, target_guess=(\"diagnosis\",\"target\",\"class\",\"label\",\"outcome\")):\n", " df = pd.read_csv(path)\n", " df.columns = [normalize_col(c) for c in df.columns]\n", "\n", " # Drop junk cols\n", " junk_like = [c for c in df.columns if c.startswith(\"unnamed\")]\n", " df = df.drop(columns=junk_like, errors=\"ignore\")\n", "\n", " df = df.drop_duplicates()\n", "\n", " # Find target\n", " target_col = None\n", " for c in target_guess:\n", " if c in df.columns:\n", " target_col = c\n", " break\n", " if target_col is None:\n", " raise ValueError(\"Target column not found.\")\n", "\n", " df = df.dropna(subset=[target_col]).copy()\n", "\n", " # Map target if M/B\n", " if df[target_col].dtype == object:\n", " if set(df[target_col].unique()).issubset({\"M\",\"B\",\"m\",\"b\",\"Malignant\",\"Benign\"}):\n", " df[target_col] = df[target_col].astype(str).str.upper().map(\n", " {\"M\":1, \"B\":0, \"MALIGNANT\":1, \"BENIGN\":0}\n", " )\n", "\n", " # Feature types\n", " numeric_cols = df.select_dtypes(include=[np.number]).columns.tolist()\n", " cat_cols = [c for c in df.columns if c not in numeric_cols]\n", "\n", " if target_col in numeric_cols: numeric_cols.remove(target_col)\n", " if target_col in cat_cols: cat_cols.remove(target_col)\n", "\n", " if numeric_cols:\n", " df[numeric_cols] = SimpleImputer(strategy=\"median\").fit_transform(df[numeric_cols])\n", " if cat_cols:\n", " df[cat_cols] = SimpleImputer(strategy=\"most_frequent\").fit_transform(df[cat_cols])\n", "\n", " # Drop id column if exists\n", " for maybe_id in [\"id\",\"patient_id\",\"index\"]:\n", " if maybe_id in df.columns:\n", " df = df.drop(columns=[maybe_id])\n", "\n", " return df, target_col, numeric_cols, cat_cols\n", "\n", "df, TARGET_COL, num_cols, cat_cols = load_and_clean(RAW_CSV_PATH)\n", "df.to_csv(CLEAN_CSV_PATH, index=False)\n", "print(\"Target column:\", TARGET_COL)\n", "print(\"Numeric features:\", len(num_cols), \" | Categorical features:\", len(cat_cols))\n", "display(df.head())" ] }, { "cell_type": "code", "execution_count": 23, "id": "f419198f-f1ba-4d11-b84a-5bfec5c83696", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Train size: (455, 30) | Test size: (114, 30)\n", "Numeric features: 30 | Categorical features: 0\n" ] } ], "source": [ "X = df.drop(columns=[TARGET_COL])\n", "y = df[TARGET_COL].astype(int)\n", "\n", "X_train, X_test, y_train, y_test = train_test_split(\n", " X, y, test_size=0.2, stratify=y, random_state=RANDOM_STATE\n", ")\n", "\n", "num_cols = X_train.select_dtypes(include=[np.number]).columns.tolist()\n", "cat_cols = [c for c in X_train.columns if c not in num_cols]\n", "\n", "numeric_transformer = Pipeline(steps=[(\"scaler\", StandardScaler())])\n", "categorical_transformer = Pipeline(steps=[(\"onehot\", OneHotEncoder(handle_unknown=\"ignore\", sparse_output=False))]) if len(cat_cols) > 0 else \"drop\"\n", "\n", "preprocess = ColumnTransformer(\n", " transformers=[\n", " (\"num\", numeric_transformer, num_cols if len(num_cols)>0 else []),\n", " (\"cat\", categorical_transformer, cat_cols if len(cat_cols)>0 else [])\n", " ],\n", " remainder=\"drop\"\n", ")\n", "print(\"Train size:\", X_train.shape, \" | Test size:\", X_test.shape)\n", "print(\"Numeric features:\", len(num_cols), \" | Categorical features:\", len(cat_cols))\n" ] }, { "cell_type": "code", "execution_count": 24, "id": "ebb9608b-5a04-4ca5-8834-ebc8dffa76f8", "metadata": {}, "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", "
modelbest_paramsaccuracyf1roc_auc
0logreg{'clf__C': 1, 'clf__penalty': 'l2', 'clf__solv...0.9736840.9638550.996032
1svm{'clf__C': 1, 'clf__gamma': 'scale', 'clf__ker...0.9736840.9629630.994709
2rf{'clf__max_depth': 10, 'clf__min_samples_split...0.9736840.9629630.992890
3knn{'clf__metric': 'euclidean', 'clf__n_neighbors...0.9385960.9113920.981812
\n", "
" ], "text/plain": [ " model best_params accuracy \\\n", "0 logreg {'clf__C': 1, 'clf__penalty': 'l2', 'clf__solv... 0.973684 \n", "1 svm {'clf__C': 1, 'clf__gamma': 'scale', 'clf__ker... 0.973684 \n", "2 rf {'clf__max_depth': 10, 'clf__min_samples_split... 0.973684 \n", "3 knn {'clf__metric': 'euclidean', 'clf__n_neighbors... 0.938596 \n", "\n", " f1 roc_auc \n", "0 0.963855 0.996032 \n", "1 0.962963 0.994709 \n", "2 0.962963 0.992890 \n", "3 0.911392 0.981812 " ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "\n", "models_and_grids = {\n", " \"logreg\": (\n", " LogisticRegression(max_iter=500, random_state=RANDOM_STATE),\n", " {\"clf__C\":[0.01,0.1,1,10], \"clf__penalty\":[\"l2\"], \"clf__solver\":[\"liblinear\",\"lbfgs\"]}\n", " ),\n", " \"svm\": (\n", " SVC(probability=True, random_state=RANDOM_STATE),\n", " {\"clf__C\":[0.1,1,10], \"clf__kernel\":[\"rbf\",\"linear\"], \"clf__gamma\":[\"scale\",\"auto\"]}\n", " ),\n", " \"rf\": (\n", " RandomForestClassifier(random_state=RANDOM_STATE),\n", " {\"clf__n_estimators\":[100,300], \"clf__max_depth\":[None,5,10], \"clf__min_samples_split\":[2,5]}\n", " ),\n", " \"knn\": (\n", " KNeighborsClassifier(),\n", " {\"clf__n_neighbors\":[3,5,7,9], \"clf__weights\":[\"uniform\",\"distance\"], \"clf__metric\":[\"euclidean\",\"manhattan\"]}\n", " )\n", "}\n", "\n", "cv = StratifiedKFold(n_splits=CV_SPLITS, shuffle=True, random_state=RANDOM_STATE)\n", "\n", "results = []\n", "best_estimators = {}\n", "\n", "for name, (model, grid) in models_and_grids.items():\n", " pipe = Pipeline([(\"prep\", preprocess), (\"clf\", model)])\n", " gs = GridSearchCV(pipe, grid, cv=cv, scoring=\"roc_auc\", n_jobs=-1, refit=True)\n", " gs.fit(X_train, y_train)\n", "\n", " y_pred = gs.predict(X_test)\n", " y_proba = gs.predict_proba(X_test)[:,1]\n", "\n", " results.append({\n", " \"model\": name,\n", " \"best_params\": gs.best_params_,\n", " \"accuracy\": accuracy_score(y_test, y_pred),\n", " \"f1\": f1_score(y_test, y_pred),\n", " \"roc_auc\": roc_auc_score(y_test, y_proba)\n", " })\n", " best_estimators[name] = gs.best_estimator_\n", "\n", "baseline_results = pd.DataFrame(results).sort_values(\"roc_auc\", ascending=False)\n", "display(baseline_results)\n" ] }, { "cell_type": "code", "execution_count": 25, "id": "3743071f-360f-434b-a7d0-833d8c60641f", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "SelectKBest features: ['concave_points_worst', 'perimeter_worst', 'radius_worst', 'concave_points_mean', 'perimeter_mean', 'area_worst', 'radius_mean', 'area_mean', 'concavity_mean', 'concavity_worst']\n", "RFE features: [np.str_('radius_mean'), np.str_('concavity_mean'), np.str_('texture_se'), np.str_('perimeter_se'), np.str_('radius_worst'), np.str_('smoothness_worst'), np.str_('compactness_worst'), np.str_('concavity_worst'), np.str_('concave_points_worst'), np.str_('symmetry_worst')]\n" ] } ], "source": [ "skb = SelectKBest(score_func=f_classif, k=10)\n", "skb.fit(X_train[num_cols], y_train)\n", "skb_features = list(pd.Series(skb.scores_, index=num_cols).nlargest(10).index)\n", "\n", "\n", "rfe = RFE(\n", " estimator=LogisticRegression(\n", " max_iter=5000, \n", " solver=\"lbfgs\",\n", " random_state=RANDOM_STATE\n", " ),\n", " n_features_to_select=10\n", ")\n", "rfe.fit(X_train[num_cols], y_train)\n", "rfe_features = list(np.array(num_cols)[rfe.support_])\n", "\n", "print(\"SelectKBest features:\", skb_features)\n", "print(\"RFE features:\", rfe_features)\n" ] }, { "cell_type": "code", "execution_count": 26, "id": "1e6ee5e2-48ba-4f0f-ae39-85b2968017e7", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\n", " Results using SelectKBest features\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", "
modelaccuracyf1roc_auc
0logreg0.9824560.9756100.998677
1svm0.9736840.9629630.998677
3knn0.9561400.9367090.996197
2rf0.9649120.9500000.992394
\n", "
" ], "text/plain": [ " model accuracy f1 roc_auc\n", "0 logreg 0.982456 0.975610 0.998677\n", "1 svm 0.973684 0.962963 0.998677\n", "3 knn 0.956140 0.936709 0.996197\n", "2 rf 0.964912 0.950000 0.992394" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stdout", "output_type": "stream", "text": [ "\n", " Results using RFE features\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", "
modelaccuracyf1roc_auc
0logreg0.9824560.9761900.997354
1svm0.9649120.9512200.996693
3knn0.9473680.9250000.995205
2rf0.9561400.9367090.992725
\n", "
" ], "text/plain": [ " model accuracy f1 roc_auc\n", "0 logreg 0.982456 0.976190 0.997354\n", "1 svm 0.964912 0.951220 0.996693\n", "3 knn 0.947368 0.925000 0.995205\n", "2 rf 0.956140 0.936709 0.992725" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from IPython.display import display\n", "\n", "def retrain_on_features(selected_features, tag):\n", " selected_features = [str(f) for f in selected_features]\n", " selected_features = [f for f in selected_features if f in X_train.columns]\n", "\n", " prep = ColumnTransformer([\n", " (\"num\", StandardScaler(), selected_features)\n", " ], remainder=\"drop\")\n", "\n", " res = []\n", " for name,(model,grid) in models_and_grids.items():\n", " pipe = Pipeline([(\"prep\", prep), (\"clf\", model)])\n", " gs = GridSearchCV(pipe, grid, cv=cv, scoring=\"roc_auc\", n_jobs=-1, refit=True)\n", " gs.fit(X_train[selected_features], y_train)\n", " y_pred = gs.predict(X_test[selected_features])\n", " y_proba = gs.predict_proba(X_test[selected_features])[:,1]\n", " res.append({\n", " \"model\": name,\n", " \"accuracy\": accuracy_score(y_test, y_pred),\n", " \"f1\": f1_score(y_test, y_pred),\n", " \"roc_auc\": roc_auc_score(y_test, y_proba)\n", " })\n", "\n", " df_res = pd.DataFrame(res).sort_values(\"roc_auc\", ascending=False)\n", " print(f\"\\n Results using {tag} features\")\n", " display(df_res) # <-- force table display\n", " return df_res\n", "\n", "skb_results = retrain_on_features(skb_features, \"SelectKBest\")\n", "rfe_results = retrain_on_features(rfe_features, \"RFE\")\n" ] }, { "cell_type": "code", "execution_count": 27, "id": "36827f2a-6c24-458b-8892-14fb9bc19f1a", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Final Project Summary\n", "----------------------------------------\n", "Target column: diagnosis\n", "Train rows: 455 | Test rows: 114\n", "\n", " Baseline Best Model:\n", " model best_params accuracy f1 roc_auc\n", "logreg {'clf__C': 1, 'clf__penalty': 'l2', 'clf__solver': 'liblinear'} 0.973684 0.963855 0.996032\n", "\n", " SelectKBest Top Features:\n", "concave_points_worst, perimeter_worst, radius_worst, concave_points_mean, perimeter_mean, area_worst, radius_mean, area_mean, concavity_mean, concavity_worst\n", "\n", " RFE Top Features:\n", "radius_mean, concavity_mean, texture_se, perimeter_se, radius_worst, smoothness_worst, compactness_worst, concavity_worst, concave_points_worst, symmetry_worst\n", "----------------------------------------\n", " Project Complete\n" ] } ], "source": [ "\n", "summary = {\n", " \"target\": TARGET_COL,\n", " \"baseline_best\": baseline_results.head(1).to_dict(orient=\"records\")[0],\n", " \"selectkbest_features\": skb_features,\n", " \"rfe_features\": rfe_features\n", "}\n", "print(\"Final Project Summary\")\n", "print(\"-\" * 40)\n", "print(f\"Target column: {TARGET_COL}\")\n", "print(f\"Train rows: {X_train.shape[0]} | Test rows: {X_test.shape[0]}\")\n", "print()\n", "print(\" Baseline Best Model:\")\n", "print(baseline_results.head(1).to_string(index=False))\n", "print()\n", "print(\" SelectKBest Top Features:\")\n", "print(\", \".join(skb_features))\n", "print()\n", "print(\" RFE Top Features:\")\n", "print(\", \".join(rfe_features))\n", "print(\"-\" * 40)\n", "print(\" Project Complete\")" ] }, { "cell_type": "code", "execution_count": null, "id": "de239aa3-2dbc-4966-b16f-d8ca10b94daa", "metadata": {}, "outputs": [], "source": [] } ], "metadata": { "kernelspec": { "display_name": "Python 3 (ipykernel)", "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": "3.13.5" } }, "nbformat": 4, "nbformat_minor": 5 }