finaltry/dano2025
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fully working spam hypot

Browse Source
This commit is contained in:
dan
2025-12-12 23:27:23 +03:00
parent ce595182b9
commit c5c10d1fcf
19 changed files with 39 additions and 3882 deletions
Download Patch File Download Diff File Expand all files Collapse all files

188
spam_hypot/01_stat_analysis.ipynb
View File

@@ -1,188 +0,0 @@
{
"cells": [
{
"cell_type": "markdown",
"id": "4d7d3347",
"metadata": {},
"source": [
"# Спам-гипотеза: плотность показов vs CTR/CR\n",
"\n",
"Цель: проверить, что высокая плотность показов на контактный день снижает CTR и CR (спам-эффект)."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "7acbd1c8",
"metadata": {},
"outputs": [],
"source": [
"import sqlite3\n",
"from pathlib import Path\n",
"import sys\n",
"import numpy as np\n",
"import pandas as pd\n",
"import seaborn as sns\n",
"import matplotlib.pyplot as plt\n",
"from scipy import stats\n",
"from sklearn.model_selection import train_test_split\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 roc_auc_score\n",
"\n",
"sns.set_theme(style=\"whitegrid\")\n",
"plt.rcParams[\"figure.figsize\"] = (10, 5)\n",
"\n",
"project_root = Path.cwd().resolve()\n",
"while not (project_root / \"preanalysis\").exists() and project_root.parent != project_root:\n",
" project_root = project_root.parent\n",
"sys.path.append(str(project_root / \"preanalysis\"))\n",
"import eda_utils as eda\n",
"\n",
"db_path = project_root / \"dataset\" / \"ds.sqlite\"\n",
"conn = sqlite3.connect(db_path)\n",
"df = pd.read_sql_query(\"select * from communications\", conn, parse_dates=[\"business_dt\"])\n",
"conn.close()\n",
"\n",
"for cols, name in [\n",
" (eda.ACTIVE_IMP_COLS, \"active_imp_total\"),\n",
" (eda.PASSIVE_IMP_COLS, \"passive_imp_total\"),\n",
" (eda.ACTIVE_CLICK_COLS, \"active_click_total\"),\n",
" (eda.PASSIVE_CLICK_COLS, \"passive_click_total\"),\n",
" (eda.ORDER_COLS, \"orders_amt_total\"),\n",
"]:\n",
" df[name] = df[cols].sum(axis=1)\n",
"\n",
"df[\"imp_total\"] = df[\"active_imp_total\"] + df[\"passive_imp_total\"]\n",
"df[\"click_total\"] = df[\"active_click_total\"] + df[\"passive_click_total\"]\n",
"\n",
"contact_days = df.groupby(\"id\")[\"business_dt\"].nunique().rename(\"contact_days\")\n",
"client = df.groupby(\"id\").agg(\n",
" {\n",
" \"imp_total\": \"sum\",\n",
" \"click_total\": \"sum\",\n",
" \"orders_amt_total\": \"sum\",\n",
" \"age\": \"median\",\n",
" \"gender_cd\": lambda s: s.mode().iat[0],\n",
" \"device_platform_cd\": lambda s: s.mode().iat[0],\n",
" }\n",
").merge(contact_days, on=\"id\", how=\"left\").reset_index()\n",
"\n",
"client[\"ctr_all\"] = eda.safe_divide(client[\"click_total\"], client[\"imp_total\"])\n",
"client[\"cr_click2order\"] = eda.safe_divide(client[\"orders_amt_total\"], client[\"click_total\"])\n",
"client[\"avg_imp_per_day\"] = eda.safe_divide(client[\"imp_total\"], client[\"contact_days\"])\n",
"client[\"high_ctr\"] = (client[\"ctr_all\"] >= client[\"ctr_all\"].quantile(0.75)).astype(int)\n",
"client[\"has_order\"] = (client[\"orders_amt_total\"] > 0).astype(int)\n"
]
},
{
"cell_type": "markdown",
"id": "94eb2d26",
"metadata": {},
"source": [
"## Базовые статистики"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "287a09b4",
"metadata": {},
"outputs": [],
"source": [
"summary = client[[\"imp_total\", \"click_total\", \"orders_amt_total\", \"contact_days\", \"avg_imp_per_day\", \"ctr_all\", \"cr_click2order\"]].describe().T\n",
"missing = client.isna().mean().sort_values(ascending=False)\n",
"summary, missing.head(10)\n"
]
},
{
"cell_type": "markdown",
"id": "10cd44b7",
"metadata": {},
"source": [
"## Корреляции и тесты\n",
"Спирмен между плотностью и CTR/CR, а также Mann–Whitney между Q1 и Q4 по плотности."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "88714a03",
"metadata": {},
"outputs": [],
"source": [
"corr_ctr = stats.spearmanr(client[\"avg_imp_per_day\"], client[\"ctr_all\"])\n",
"corr_cr = stats.spearmanr(client[\"avg_imp_per_day\"], client[\"cr_click2order\"])\n",
"q1 = client[\"avg_imp_per_day\"].quantile(0.25)\n",
"q4 = client[\"avg_imp_per_day\"].quantile(0.75)\n",
"low = client.loc[client[\"avg_imp_per_day\"] <= q1, \"ctr_all\"].dropna()\n",
"high = client.loc[client[\"avg_imp_per_day\"] >= q4, \"ctr_all\"].dropna()\n",
"wu = stats.mannwhitneyu(low, high, alternative=\"greater\")\n",
"{ \"spearman_ctr\": corr_ctr, \"spearman_cr\": corr_cr, \"mw_low_gt_high\": wu }\n"
]
},
{
"cell_type": "markdown",
"id": "20d492fa",
"metadata": {},
"source": [
"bins = pd.qcut(client[\"avg_imp_per_day\"], 10, duplicates=\"drop\")\n",
"stats_bin = client.groupby(bins, observed=False).agg(\n",
" ctr_all=(\"ctr_all\", \"median\"),\n",
" cr_click2order=(\"cr_click2order\", \"median\"),\n",
" avg_imp_per_day=(\"avg_imp_per_day\", \"median\"),\n",
").reset_index()\n",
"stats_bin[\"bin_label\"] = stats_bin[\"avg_imp_per_day\"].round(2).astype(str)\n",
"fig, ax1 = plt.subplots(figsize=(12, 5))\n",
"ax2 = ax1.twinx()\n",
"sns.lineplot(data=stats_bin, x=\"bin_label\", y=\"ctr_all\", marker=\"o\", ax=ax1, color=\"#4c72b0\", label=\"CTR\")\n",
"sns.lineplot(data=stats_bin, x=\"bin_label\", y=\"cr_click2order\", marker=\"o\", ax=ax2, color=\"#c44e52\", label=\"CR\")\n",
"ax1.set_ylabel(\"CTR\")\n",
"ax2.set_ylabel(\"CR click→order\")\n",
"plt.xticks(rotation=35)\n",
"ax1.set_title(\"CTR и CR по децилям avg_imp_per_day\")\n",
"fig.tight_layout()\n",
"plt.show()\n",
"stats_bin[[\"bin_label\", \"ctr_all\", \"cr_click2order\"]]\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "943f0d4b",
"metadata": {},
"outputs": [],
"source": [
"bins = pd.qcut(client[\"avg_imp_per_day\"], 10, duplicates=\"drop\")\n",
"stats_bin = client.groupby(bins).agg({\"ctr_all\": \"median\", \"cr_click2order\": \"median\", \"avg_imp_per_day\": \"median\"}).reset_index()\n",
"stats_bin[\"bin_label\"] = stats_bin[\"avg_imp_per_day\"].round(2).astype(str)\n",
"fig, ax1 = plt.subplots(figsize=(12, 5))\n",
"ax2 = ax1.twinx()\n",
"sns.lineplot(data=stats_bin, x=\"bin_label\", y=\"ctr_all\", marker=\"o\", ax=ax1, color=\"#4c72b0\", label=\"CTR\")\n",
"sns.lineplot(data=stats_bin, x=\"bin_label\", y=\"cr_click2order\", marker=\"o\", ax=ax2, color=\"#c44e52\", label=\"CR\")\n",
"ax1.set_ylabel(\"CTR\")\n",
"ax2.set_ylabel(\"CR click→order\")\n",
"plt.xticks(rotation=35)\n",
"ax1.set_title(\"CTR и CR по децилям avg_imp_per_day\")\n",
"fig.tight_layout()\n",
"plt.show()\n",
"stats_bin[[\"bin_label\", \"ctr_all\", \"cr_click2order\"]]\n"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python",
"version": "3.13"
}
},
"nbformat": 4,
"nbformat_minor": 5
}

161
spam_hypot/02_models.ipynb
View File

@@ -1,161 +0,0 @@
{
"cells": [
{
"cell_type": "markdown",
"id": "7254b4c1",
"metadata": {},
"source": [
"# Спам-гипотеза: сравнение моделей\n",
"\n",
"Target: `high_ctr` (верхний квартиль CTR)."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "c7f54168",
"metadata": {},
"outputs": [],
"source": [
"import sqlite3\n",
"from pathlib import Path\n",
"import sys\n",
"import numpy as np\n",
"import pandas as pd\n",
"import seaborn as sns\n",
"import matplotlib.pyplot as plt\n",
"from scipy import stats\n",
"from sklearn.model_selection import train_test_split\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 roc_auc_score\n",
"\n",
"sns.set_theme(style=\"whitegrid\")\n",
"plt.rcParams[\"figure.figsize\"] = (10, 5)\n",
"\n",
"project_root = Path.cwd().resolve()\n",
"while not (project_root / \"preanalysis\").exists() and project_root.parent != project_root:\n",
" project_root = project_root.parent\n",
"sys.path.append(str(project_root / \"preanalysis\"))\n",
"import eda_utils as eda\n",
"\n",
"db_path = project_root / \"dataset\" / \"ds.sqlite\"\n",
"conn = sqlite3.connect(db_path)\n",
"df = pd.read_sql_query(\"select * from communications\", conn, parse_dates=[\"business_dt\"])\n",
"conn.close()\n",
"\n",
"for cols, name in [\n",
" (eda.ACTIVE_IMP_COLS, \"active_imp_total\"),\n",
" (eda.PASSIVE_IMP_COLS, \"passive_imp_total\"),\n",
" (eda.ACTIVE_CLICK_COLS, \"active_click_total\"),\n",
" (eda.PASSIVE_CLICK_COLS, \"passive_click_total\"),\n",
" (eda.ORDER_COLS, \"orders_amt_total\"),\n",
"]:\n",
" df[name] = df[cols].sum(axis=1)\n",
"\n",
"df[\"imp_total\"] = df[\"active_imp_total\"] + df[\"passive_imp_total\"]\n",
"df[\"click_total\"] = df[\"active_click_total\"] + df[\"passive_click_total\"]\n",
"\n",
"contact_days = df.groupby(\"id\")[\"business_dt\"].nunique().rename(\"contact_days\")\n",
"client = df.groupby(\"id\").agg(\n",
" {\n",
" \"imp_total\": \"sum\",\n",
" \"click_total\": \"sum\",\n",
" \"orders_amt_total\": \"sum\",\n",
" \"age\": \"median\",\n",
" \"gender_cd\": lambda s: s.mode().iat[0],\n",
" \"device_platform_cd\": lambda s: s.mode().iat[0],\n",
" }\n",
").merge(contact_days, on=\"id\", how=\"left\").reset_index()\n",
"\n",
"client[\"ctr_all\"] = eda.safe_divide(client[\"click_total\"], client[\"imp_total\"])\n",
"client[\"cr_click2order\"] = eda.safe_divide(client[\"orders_amt_total\"], client[\"click_total\"])\n",
"client[\"avg_imp_per_day\"] = eda.safe_divide(client[\"imp_total\"], client[\"contact_days\"])\n",
"client[\"high_ctr\"] = (client[\"ctr_all\"] >= client[\"ctr_all\"].quantile(0.75)).astype(int)\n",
"client[\"has_order\"] = (client[\"orders_amt_total\"] > 0).astype(int)\n"
]
},
{
"cell_type": "markdown",
"id": "21786c63",
"metadata": {},
"source": [
"## Модели: Logistic Regression vs GradientBoosting"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "dc8dbc94",
"metadata": {},
"outputs": [],
"source": [
"X = client[[\"avg_imp_per_day\", \"imp_total\", \"click_total\", \"age\", \"gender_cd\", \"device_platform_cd\"]]\n",
"X = X.copy()\n",
"X[\"gender_cd\"] = eda.normalize_gender(X[\"gender_cd\"])\n",
"X[\"device_platform_cd\"] = eda.normalize_device(X[\"device_platform_cd\"])\n",
"y = client[\"high_ctr\"]\n",
"\n",
"num_cols = [\"avg_imp_per_day\", \"imp_total\", \"click_total\", \"age\"]\n",
"cat_cols = [\"gender_cd\", \"device_platform_cd\"]\n",
"pre = ColumnTransformer([\n",
" (\"num\", Pipeline([(\"imputer\", SimpleImputer(strategy=\"median\")), (\"scaler\", StandardScaler())]), num_cols),\n",
" (\"cat\", OneHotEncoder(handle_unknown=\"ignore\"), cat_cols),\n",
"])\n",
"\n",
"log_reg = Pipeline([(\"pre\", pre), (\"clf\", LogisticRegression(max_iter=1000))])\n",
"gb = Pipeline([(\"pre\", pre), (\"clf\", GradientBoostingClassifier(random_state=42))])\n",
"\n",
"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)\n",
"res = {}\n",
"for name, model in [(\"log_reg\", log_reg), (\"gb\", gb)]:\n",
" model.fit(X_train, y_train)\n",
" proba = model.predict_proba(X_test)[:, 1]\n",
" res[name] = roc_auc_score(y_test, proba)\n",
"res\n"
]
},
{
"cell_type": "markdown",
"id": "203acf70",
"metadata": {},
"source": [
"## Важности признаков (GradientBoosting)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "3eac9e17",
"metadata": {},
"outputs": [],
"source": [
"gb_model = gb\n",
"feat_names = gb_model.named_steps[\"pre\"].get_feature_names_out()\n",
"importances = gb_model.named_steps[\"clf\"].feature_importances_\n",
"imp_df = pd.DataFrame({\"feature\": feat_names, \"importance\": importances}).sort_values(\"importance\", ascending=False)\n",
"plt.figure(figsize=(8, 5))\n",
"sns.barplot(data=imp_df.head(15), x=\"importance\", y=\"feature\", palette=\"viridis\")\n",
"plt.title(\"Top-15 feature importances (GB)\")\n",
"plt.tight_layout()\n",
"plt.show()\n",
"imp_df.head(15)\n"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python",
"version": "3.13"
}
},
"nbformat": 4,
"nbformat_minor": 5
}

206
spam_hypot/03_best_model.ipynb
View File

@@ -1,206 +0,0 @@
{
"cells": [
{
"cell_type": "markdown",
"id": "d88bf2d8",
"metadata": {},
"source": [
"# Спам-гипотеза: лучшая модель и визуализации\n",
"\n",
"Используем GradientBoostingClassifier (лучше логрега по AUC) для подтверждения гипотезы."
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "87f3f728",
"metadata": {},
"outputs": [],
"source": [
"import sqlite3\n",
"from pathlib import Path\n",
"import sys\n",
"import numpy as np\n",
"import pandas as pd\n",
"import seaborn as sns\n",
"import matplotlib.pyplot as plt\n",
"from scipy import stats\n",
"from sklearn.model_selection import train_test_split\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 roc_auc_score\n",
"\n",
"sns.set_theme(style=\"whitegrid\")\n",
"plt.rcParams[\"figure.figsize\"] = (10, 5)\n",
"\n",
"project_root = Path.cwd().resolve()\n",
"while not (project_root / \"preanalysis\").exists() and project_root.parent != project_root:\n",
" project_root = project_root.parent\n",
"sys.path.append(str(project_root / \"preanalysis\"))\n",
"import eda_utils as eda\n",
"\n",
"db_path = project_root / \"dataset\" / \"ds.sqlite\"\n",
"conn = sqlite3.connect(db_path)\n",
"df = pd.read_sql_query(\"select * from communications\", conn, parse_dates=[\"business_dt\"])\n",
"conn.close()\n",
"\n",
"for cols, name in [\n",
" (eda.ACTIVE_IMP_COLS, \"active_imp_total\"),\n",
" (eda.PASSIVE_IMP_COLS, \"passive_imp_total\"),\n",
" (eda.ACTIVE_CLICK_COLS, \"active_click_total\"),\n",
" (eda.PASSIVE_CLICK_COLS, \"passive_click_total\"),\n",
" (eda.ORDER_COLS, \"orders_amt_total\"),\n",
"]:\n",
" df[name] = df[cols].sum(axis=1)\n",
"\n",
"df[\"imp_total\"] = df[\"active_imp_total\"] + df[\"passive_imp_total\"]\n",
"df[\"click_total\"] = df[\"active_click_total\"] + df[\"passive_click_total\"]\n",
"\n",
"contact_days = df.groupby(\"id\")[\"business_dt\"].nunique().rename(\"contact_days\")\n",
"client = df.groupby(\"id\").agg(\n",
" {\n",
" \"imp_total\": \"sum\",\n",
" \"click_total\": \"sum\",\n",
" \"orders_amt_total\": \"sum\",\n",
" \"age\": \"median\",\n",
" \"gender_cd\": lambda s: s.mode().iat[0],\n",
" \"device_platform_cd\": lambda s: s.mode().iat[0],\n",
" }\n",
").merge(contact_days, on=\"id\", how=\"left\").reset_index()\n",
"\n",
"client[\"ctr_all\"] = eda.safe_divide(client[\"click_total\"], client[\"imp_total\"])\n",
"client[\"cr_click2order\"] = eda.safe_divide(client[\"orders_amt_total\"], client[\"click_total\"])\n",
"client[\"avg_imp_per_day\"] = eda.safe_divide(client[\"imp_total\"], client[\"contact_days\"])\n",
"client[\"high_ctr\"] = (client[\"ctr_all\"] >= client[\"ctr_all\"].quantile(0.75)).astype(int)\n",
"client[\"has_order\"] = (client[\"orders_amt_total\"] > 0).astype(int)\n"
]
},
{
"cell_type": "markdown",
"id": "17da010c",
"metadata": {},
"source": [
"## Обучение лучшей модели"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "81433d7e",
"metadata": {},
"outputs": [],
"source": [
"X = client[[\"avg_imp_per_day\", \"imp_total\", \"click_total\", \"age\", \"gender_cd\", \"device_platform_cd\"]]\n",
"X = X.copy()\n",
"X[\"gender_cd\"] = eda.normalize_gender(X[\"gender_cd\"])\n",
"X[\"device_platform_cd\"] = eda.normalize_device(X[\"device_platform_cd\"])\n",
"y = client[\"high_ctr\"]\n",
"\n",
"num_cols = [\"avg_imp_per_day\", \"imp_total\", \"click_total\", \"age\"]\n",
"cat_cols = [\"gender_cd\", \"device_platform_cd\"]\n",
"pre = ColumnTransformer([\n",
" (\"num\", Pipeline([(\"imputer\", SimpleImputer(strategy=\"median\")), (\"scaler\", StandardScaler())]), num_cols),\n",
" (\"cat\", OneHotEncoder(handle_unknown=\"ignore\"), cat_cols),\n",
"])\n",
"\n",
"best = Pipeline([(\"pre\", pre), (\"clf\", GradientBoostingClassifier(random_state=42))])\n",
"X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)\n",
"best.fit(X_train, y_train)\n",
"proba = best.predict_proba(X_test)[:, 1]\n",
"auc = roc_auc_score(y_test, proba)\n",
"auc\n"
]
},
{
"cell_type": "markdown",
"id": "63f4db9b",
"metadata": {},
"source": [
"## Прогноз vs плотность показов"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "f48584b5",
"metadata": {},
"outputs": [],
"source": [
"grid = pd.DataFrame({\"avg_imp_per_day\": np.linspace(client[\"avg_imp_per_day\"].min(), client[\"avg_imp_per_day\"].max(), 50)})\n",
"base = client.median(numeric_only=True)\n",
"base_gender = client[\"gender_cd\"].mode().iat[0]\n",
"base_device = client[\"device_platform_cd\"].mode().iat[0]\n",
"grid[\"imp_total\"] = base[\"imp_total\"]\n",
"grid[\"click_total\"] = base[\"click_total\"]\n",
"grid[\"age\"] = base[\"age\"]\n",
"grid[\"gender_cd\"] = base_gender\n",
"grid[\"device_platform_cd\"] = base_device\n",
"proba_grid = best.predict_proba(grid)[:, 1]\n",
"plt.figure(figsize=(10, 4))\n",
"plt.plot(grid[\"avg_imp_per_day\"], proba_grid, marker=\"o\")\n",
"plt.xlabel(\"avg_imp_per_day\")\n",
"plt.ylabel(\"P(high CTR)\")\n",
"plt.title(\"Предсказанная вероятность высокого CTR vs плотность показов\")\n",
"plt.tight_layout()\n",
"plt.show()\n"
]
},
{
"cell_type": "markdown",
"id": "32f73b44",
"metadata": {},
"source": [
"## График CTR и CR по тонким бинам (две оси)"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "bb4d0190",
"metadata": {},
"outputs": [],
"source": [
"bins = pd.qcut(client[\"avg_imp_per_day\"], 15, duplicates=\"drop\")\n",
"stats_bin = client.groupby(bins).agg({\"ctr_all\": \"median\", \"cr_click2order\": \"median\", \"avg_imp_per_day\": \"median\"}).reset_index()\n",
"stats_bin[\"bin_label\"] = stats_bin[\"avg_imp_per_day\"].round(2).astype(str)\n",
"fig, ax1 = plt.subplots(figsize=(12, 5))\n",
"ax2 = ax1.twinx()\n",
"ax1.plot(stats_bin[\"bin_label\"], stats_bin[\"ctr_all\"], marker=\"o\", color=\"#4c72b0\", label=\"CTR\")\n",
"ax2.plot(stats_bin[\"bin_label\"], stats_bin[\"cr_click2order\"], marker=\"s\", color=\"#c44e52\", label=\"CR\")\n",
"ax1.set_ylabel(\"CTR\")\n",
"ax2.set_ylabel(\"CR click→order\")\n",
"ax1.set_xlabel(\"avg_imp_per_day bins\")\n",
"plt.xticks(rotation=35)\n",
"ax1.set_title(\"CTR и CR по 15 бинам avg_imp_per_day\")\n",
"fig.tight_layout()\n",
"plt.show()\n"
]
},
{
"cell_type": "markdown",
"id": "ebb2ca5e",
"metadata": {},
"source": [
"## Вывод\n",
"- AUC модели GradientBoosting > логрега; `avg_imp_per_day` ключевой драйвер: рост плотности снижает шанс попасть в верхний квартиль CTR.\n",
"- Биновые графики показывают монотонное падение CTR и CR при росте avg_imp_per_day.\n",
"- Гипотеза о спам-эффекте подтверждается как статистически, так и по ML-модели."
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"name": "python",
"version": "3.13"
}
},
"nbformat": 4,
"nbformat_minor": 5
}

BIN
spam_hypot/best_bins.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 109 KiB

After

Width:  |  Height:  |  Size: 100 KiB

BIN
spam_hypot/best_model_prob.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 45 KiB

After

Width:  |  Height:  |  Size: 41 KiB

48
spam_hypot/model_compare.py
View File

@@ -46,32 +46,62 @@ client = (
.merge(contact_days, on="id", how="left")
.reset_index()
)
# ... всё как у тебя до расчёта client["ctr_all"] включительно
client["ctr_all"] = eda.safe_divide(client["click_total"], client["imp_total"])
client["avg_imp_per_day"] = eda.safe_divide(client["imp_total"], client["contact_days"])
client["high_ctr"] = (client["ctr_all"] >= client["ctr_all"].quantile(0.75)).astype(int)
X = client[["avg_imp_per_day", "imp_total", "click_total", "age", "gender_cd", "device_platform_cd"]]
X = X.copy()
X["gender_cd"] = eda.normalize_gender(X["gender_cd"])
X["device_platform_cd"] = eda.normalize_device(X["device_platform_cd"])
y = client["high_ctr"]
# --- SPLIT СНАЧАЛА, ТАРГЕТ ПОТОМ ---
train_idx, test_idx = train_test_split(
client.index, test_size=0.2, random_state=42
)
num_cols = ["avg_imp_per_day", "imp_total", "click_total", "age"]
train = client.loc[train_idx].copy()
test = client.loc[test_idx].copy()
thr = train["ctr_all"].quantile(0.75) # порог только по train
train["high_ctr"] = (train["ctr_all"] >= thr).astype(int)
test["high_ctr"] = (test["ctr_all"] >= thr).astype(int)
# --- ФИЧИ БЕЗ click_total (иначе это чит) ---
X_train = train[[
"avg_imp_per_day", "imp_total", "contact_days", # можно оставить
"age", "gender_cd", "device_platform_cd"
]].copy()
X_test = test[[
"avg_imp_per_day", "imp_total", "contact_days",
"age", "gender_cd", "device_platform_cd"
]].copy()
X_train["gender_cd"] = eda.normalize_gender(X_train["gender_cd"])
X_train["device_platform_cd"] = eda.normalize_device(X_train["device_platform_cd"])
X_test["gender_cd"] = eda.normalize_gender(X_test["gender_cd"])
X_test["device_platform_cd"] = eda.normalize_device(X_test["device_platform_cd"])
y_train = train["high_ctr"]
y_test = test["high_ctr"]
num_cols = ["avg_imp_per_day", "imp_total", "contact_days", "age"]
cat_cols = ["gender_cd", "device_platform_cd"]
pre = ColumnTransformer([
("num", Pipeline([("imputer", SimpleImputer(strategy="median")), ("scaler", StandardScaler())]), num_cols),
("num", Pipeline([
("imputer", SimpleImputer(strategy="median")),
("scaler", StandardScaler())
]), num_cols),
("cat", OneHotEncoder(handle_unknown="ignore"), cat_cols),
])
log_reg = Pipeline([("pre", pre), ("clf", LogisticRegression(max_iter=1000))])
gb = Pipeline([("pre", pre), ("clf", GradientBoostingClassifier(random_state=42))])
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42, stratify=y)
results = {}
for name, model in [("log_reg", log_reg), ("gb", gb)]:
model.fit(X_train, y_train)
proba = model.predict_proba(X_test)[:, 1]
results[name] = roc_auc_score(y_test, proba)
print("CTR threshold (train 0.75q):", thr)
print("AUC results:", results)
imp = gb.named_steps["clf"].feature_importances_

BIN
spam_hypot/stat_bins.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 97 KiB

After

Width:  |  Height:  |  Size: 87 KiB