{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [], "source": [ "#! pip install tensorflow\n", "#! pip install scikit-learn\n", "\n", "\n" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [], "source": [ "import pandas as pd\n", "import numpy as np\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns \n" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [], "source": [ "train_df = pd.read_csv(\"train.csv\")" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "0 poznan.wuoz.gov.pl\n", "1 vill.okawa.kochi.jp\n", "2 nationalfinance.co.om\n", "3 town.ozora.hokkaido.jp\n", "4 open24.ie-news.irish/online/Login\n", " ... \n", "6995051 ddht.co.kr\n", "6995052 www.upstartepoxy.com\n", "6995053 employeesalaryschedule70.000webhostapp.com/adb...\n", "6995054 dekalbtool.com\n", "6995055 helpinganimals.com\n", "Name: URL_clean, Length: 6995056, dtype: object" ] }, "execution_count": 4, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# 대괄호 [ ] 제거\n", "train_df['URL_clean'] = train_df['URL'].str.replace(r'[\\[\\]]', '', regex=True)\n", "train_df[\"URL_clean\"]" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
| \n", " | ID | \n", "URL | \n", "label | \n", "URL_clean | \n", "
|---|---|---|---|---|
| 0 | \n", "TRAIN_0000000 | \n", "poznan[.]wuoz[.]gov[.]pl | \n", "0 | \n", "poznan.wuoz.gov.pl | \n", "
| 1 | \n", "TRAIN_0000001 | \n", "vill[.]okawa[.]kochi[.]jp | \n", "0 | \n", "vill.okawa.kochi.jp | \n", "
| 2 | \n", "TRAIN_0000002 | \n", "nationalfinance[.]co[.]om | \n", "0 | \n", "nationalfinance.co.om | \n", "
| 3 | \n", "TRAIN_0000003 | \n", "town[.]ozora[.]hokkaido[.]jp | \n", "0 | \n", "town.ozora.hokkaido.jp | \n", "
| 4 | \n", "TRAIN_0000004 | \n", "open24[.]ie-news[.]irish/online/Login | \n", "1 | \n", "open24.ie-news.irish/online/Login | \n", "
| ... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "
| 6995051 | \n", "TRAIN_6995051 | \n", "ddht[.]co[.]kr | \n", "0 | \n", "ddht.co.kr | \n", "
| 6995052 | \n", "TRAIN_6995052 | \n", "www[.]upstartepoxy[.]com | \n", "0 | \n", "www.upstartepoxy.com | \n", "
| 6995053 | \n", "TRAIN_6995053 | \n", "employeesalaryschedule70[.]000webhostapp[.]com... | \n", "1 | \n", "employeesalaryschedule70.000webhostapp.com/adb... | \n", "
| 6995054 | \n", "TRAIN_6995054 | \n", "dekalbtool[.]com | \n", "0 | \n", "dekalbtool.com | \n", "
| 6995055 | \n", "TRAIN_6995055 | \n", "helpinganimals[.]com | \n", "0 | \n", "helpinganimals.com | \n", "
6995056 rows × 4 columns
\n", "| \n", " | ID | \n", "URL | \n", "label | \n", "URL_clean | \n", "digit_count | \n", "dash_count | \n", "underscore_count | \n", "percent_count | \n", "equal_count | \n", "question_count | \n", "... | \n", "count_of_double_slash | \n", "has_www | \n", "suspicious_word_count | \n", "path_depth | \n", "has_long_digit_sequence | \n", "has_multiple_dash | \n", "has_https | \n", "count_of_exclamation | \n", "count_of_dot | \n", "ends_with_common_extension | \n", "
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | \n", "TRAIN_0000000 | \n", "poznan[.]wuoz[.]gov[.]pl | \n", "0 | \n", "poznan.wuoz.gov.pl | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "3 | \n", "False | \n", "
| 1 | \n", "TRAIN_0000001 | \n", "vill[.]okawa[.]kochi[.]jp | \n", "0 | \n", "vill.okawa.kochi.jp | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "3 | \n", "False | \n", "
| 2 | \n", "TRAIN_0000002 | \n", "nationalfinance[.]co[.]om | \n", "0 | \n", "nationalfinance.co.om | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "
| 3 | \n", "TRAIN_0000003 | \n", "town[.]ozora[.]hokkaido[.]jp | \n", "0 | \n", "town.ozora.hokkaido.jp | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "3 | \n", "False | \n", "
| 4 | \n", "TRAIN_0000004 | \n", "open24[.]ie-news[.]irish/online/Login | \n", "1 | \n", "open24.ie-news.irish/online/Login | \n", "2 | \n", "1 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "1 | \n", "0 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "
| ... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "
| 6995051 | \n", "TRAIN_6995051 | \n", "ddht[.]co[.]kr | \n", "0 | \n", "ddht.co.kr | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "
| 6995052 | \n", "TRAIN_6995052 | \n", "www[.]upstartepoxy[.]com | \n", "0 | \n", "www.upstartepoxy.com | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "True | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "
| 6995053 | \n", "TRAIN_6995053 | \n", "employeesalaryschedule70[.]000webhostapp[.]com... | \n", "1 | \n", "employeesalaryschedule70.000webhostapp.com/adb... | \n", "5 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "0 | \n", "0 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "
| 6995054 | \n", "TRAIN_6995054 | \n", "dekalbtool[.]com | \n", "0 | \n", "dekalbtool.com | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "1 | \n", "False | \n", "
| 6995055 | \n", "TRAIN_6995055 | \n", "helpinganimals[.]com | \n", "0 | \n", "helpinganimals.com | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "0 | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "1 | \n", "False | \n", "
6995056 rows × 23 columns
\n", "| \n", " | ID | \n", "URL | \n", "label | \n", "URL_clean | \n", "digit_count | \n", "dash_count | \n", "underscore_count | \n", "percent_count | \n", "equal_count | \n", "question_count | \n", "... | \n", "has_www | \n", "suspicious_word_count | \n", "path_depth | \n", "has_long_digit_sequence | \n", "has_multiple_dash | \n", "has_https | \n", "count_of_exclamation | \n", "count_of_dot | \n", "ends_with_common_extension | \n", "url_entropy | \n", "
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 0 | \n", "TRAIN_0000000 | \n", "poznan[.]wuoz[.]gov[.]pl | \n", "0 | \n", "poznan.wuoz.gov.pl | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "3 | \n", "False | \n", "3.308271 | \n", "
| 1 | \n", "TRAIN_0000001 | \n", "vill[.]okawa[.]kochi[.]jp | \n", "0 | \n", "vill.okawa.kochi.jp | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "3 | \n", "False | \n", "3.471354 | \n", "
| 2 | \n", "TRAIN_0000002 | \n", "nationalfinance[.]co[.]om | \n", "0 | \n", "nationalfinance.co.om | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "3.272804 | \n", "
| 3 | \n", "TRAIN_0000003 | \n", "town[.]ozora[.]hokkaido[.]jp | \n", "0 | \n", "town.ozora.hokkaido.jp | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "3 | \n", "False | \n", "3.533771 | \n", "
| 4 | \n", "TRAIN_0000004 | \n", "open24[.]ie-news[.]irish/online/Login | \n", "1 | \n", "open24.ie-news.irish/online/Login | \n", "2 | \n", "1 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "1 | \n", "0 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "3.772450 | \n", "
| ... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "... | \n", "
| 6995051 | \n", "TRAIN_6995051 | \n", "ddht[.]co[.]kr | \n", "0 | \n", "ddht.co.kr | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "2.921928 | \n", "
| 6995052 | \n", "TRAIN_6995052 | \n", "www[.]upstartepoxy[.]com | \n", "0 | \n", "www.upstartepoxy.com | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "True | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "3.684184 | \n", "
| 6995053 | \n", "TRAIN_6995053 | \n", "employeesalaryschedule70[.]000webhostapp[.]com... | \n", "1 | \n", "employeesalaryschedule70.000webhostapp.com/adb... | \n", "5 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "0 | \n", "0 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "2 | \n", "False | \n", "4.130881 | \n", "
| 6995054 | \n", "TRAIN_6995054 | \n", "dekalbtool[.]com | \n", "0 | \n", "dekalbtool.com | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "1 | \n", "False | \n", "3.324863 | \n", "
| 6995055 | \n", "TRAIN_6995055 | \n", "helpinganimals[.]com | \n", "0 | \n", "helpinganimals.com | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "0 | \n", "... | \n", "False | \n", "0 | \n", "-2 | \n", "False | \n", "False | \n", "False | \n", "0 | \n", "1 | \n", "False | \n", "3.614369 | \n", "
6995056 rows × 24 columns
\n", "Model: \"sequential\"\n",
"\n"
],
"text/plain": [
"\u001b[1mModel: \"sequential\"\u001b[0m\n"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
"┃ Layer (type) ┃ Output Shape ┃ Param # ┃\n",
"┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
"│ dense (Dense) │ (None, 64) │ 1,344 │\n",
"├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
"│ dense_1 (Dense) │ (None, 32) │ 2,080 │\n",
"├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
"│ dense_2 (Dense) │ (None, 1) │ 33 │\n",
"└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
"\n"
],
"text/plain": [
"┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┓\n",
"┃\u001b[1m \u001b[0m\u001b[1mLayer (type) \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
"┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━┩\n",
"│ dense (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m64\u001b[0m) │ \u001b[38;5;34m1,344\u001b[0m │\n",
"├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
"│ dense_1 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m) │ \u001b[38;5;34m2,080\u001b[0m │\n",
"├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
"│ dense_2 (\u001b[38;5;33mDense\u001b[0m) │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m) │ \u001b[38;5;34m33\u001b[0m │\n",
"└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"Total params: 3,457 (13.50 KB)\n", "\n" ], "text/plain": [ "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m3,457\u001b[0m (13.50 KB)\n" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
Trainable params: 3,457 (13.50 KB)\n", "\n" ], "text/plain": [ "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m3,457\u001b[0m (13.50 KB)\n" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
Non-trainable params: 0 (0.00 B)\n", "\n" ], "text/plain": [ "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from tensorflow.keras.models import Sequential\n", "from tensorflow.keras.layers import Dense\n", "\n", "# 모델 구성 \n", "model = Sequential()\n", "model.add(Dense(64, input_dim=X_train_scaled.shape[1], activation='relu')) # 첫번째 은닉층 노드 64개 input_dim=X_train.shape[1]: 입력피처 개수 설정 \n", "model.add(Dense(32, activation='relu')) # 이진분류 두번쨰 은닉층 \n", "model.add(Dense(1, activation='sigmoid'))\n", "\n", "# 모델 컴파일\n", "from tensorflow.keras.optimizers import Adam\n", "\n", "optimizer = Adam(learning_rate=0.001)\n", "model.compile(optimizer=optimizer, loss='binary_crossentropy', metrics=['accuracy'])\n", "\n", "model.summary()\n" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [], "source": [ "# EarlyStopping 콜백을 사용하여 검증 손실이 개선되지 않으면 학습을 멈추도록 설정\n", "from tensorflow.keras.callbacks import EarlyStopping\n", "early_stopping = EarlyStopping(monitor='val_loss', patience=5)" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Epoch 1/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m19s\u001b[0m 2ms/step - accuracy: 0.8508 - loss: 0.3584 - val_accuracy: 0.8753 - val_loss: 0.3081\n", "Epoch 2/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8744 - loss: 0.3074 - val_accuracy: 0.8777 - val_loss: 0.3025\n", "Epoch 3/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8780 - loss: 0.3027 - val_accuracy: 0.8784 - val_loss: 0.3006\n", "Epoch 4/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8797 - loss: 0.2995 - val_accuracy: 0.8815 - val_loss: 0.2969\n", "Epoch 5/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8810 - loss: 0.2974 - val_accuracy: 0.8832 - val_loss: 0.2938\n", "Epoch 6/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8819 - loss: 0.2956 - val_accuracy: 0.8840 - val_loss: 0.2928\n", "Epoch 7/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8827 - loss: 0.2941 - val_accuracy: 0.8846 - val_loss: 0.2921\n", "Epoch 8/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8834 - loss: 0.2932 - val_accuracy: 0.8851 - val_loss: 0.2904\n", "Epoch 9/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8837 - loss: 0.2921 - val_accuracy: 0.8842 - val_loss: 0.2914\n", "Epoch 10/10\n", "\u001b[1m8744/8744\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m17s\u001b[0m 2ms/step - accuracy: 0.8842 - loss: 0.2919 - val_accuracy: 0.8856 - val_loss: 0.2898\n", "\u001b[1m43720/43720\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m50s\u001b[0m 1ms/step - accuracy: 0.8853 - loss: 0.2897\n", "Validation Accuracy: 0.8854\n" ] } ], "source": [ "# from sklearn.utils import class_weight\n", "\n", "\n", "# # 클래스 가중치 계산 (불균형 보정)\n", "# weights = class_weight.compute_class_weight(\n", "# class_weight='balanced',\n", "# classes=np.unique(y_train),\n", "# y=y_train\n", "# )\n", "# class_weights = dict(zip(np.unique(y_train), weights))\n", "\n", "# 모델 학습 (스케일링된 데이터 사용 + 클래스 가중치 적용)\n", "history = model.fit(\n", " X_train_scaled, y_train, # ← 스케일링된 훈련 데이터\n", " epochs=10,\n", " batch_size=512,\n", " validation_data = (X_val_scaled, y_val),\n", " # class_weight = class_weights\n", ")\n", "\n", "# 모델 평가\n", "from sklearn.metrics import accuracy_score\n", "val_loss, val_accuracy = model.evaluate(X_test_scaled, y_test)\n", "print(f'Validation Accuracy: {val_accuracy:.4f}')" ] }, { "cell_type": "code", "execution_count": 35, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "\u001b[1m34976/34976\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m26s\u001b[0m 731us/step\n" ] } ], "source": [ "from sklearn.metrics import f1_score, accuracy_score\n", "\n", "y_val_proba = model.predict(X_val_scaled).ravel() \n", "\n", "f1_scores = []\n", "accuracies = []\n", "thresholds = []\n", "\n", "for t in np.arange(0.0, 1.0, 0.01):\n", " y_pred = (y_val_proba > t).astype(int)\n", " f1 = f1_score(y_val, y_pred, zero_division=0)\n", " acc = accuracy_score(y_val, y_pred)\n", "\n", " thresholds.append(t)\n", " f1_scores.append(f1)\n", " accuracies.append(acc)\n" ] }, { "cell_type": "code", "execution_count": 38, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Best Threshold: 0.34\n", "Best F1 Score: 0.7249\n", "Accuracy at Best Threshold: 0.8779\n" ] } ], "source": [ "best_idx = np.argmax(f1_scores)\n", "best_threshold = thresholds[best_idx]\n", "best_f1 = f1_scores[best_idx]\n", "best_acc = accuracies[best_idx]\n", "\n", "print(f\"Best Threshold: {best_threshold:.2f}\")\n", "print(f\"Best F1 Score: {best_f1:.4f}\")\n", "print(f\"Accuracy at Best Threshold: {best_acc:.4f}\")\n" ] }, { "cell_type": "code", "execution_count": 39, "metadata": {}, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "