dml

https://github.com/r26raihan/dml
Science Score: 44.0%

This score indicates how likely this project is to be science-related based on various indicators:
✓
CITATION.cff file
Found CITATION.cff file
✓
codemeta.json file
Found codemeta.json file
✓
.zenodo.json file
Found .zenodo.json file
○
DOI references
○
Academic publication links
○
Academic email domains
○
Institutional organization owner
○
JOSS paper metadata
○
Scientific vocabulary similarity
Low similarity (10.6%) to scientific vocabulary
Last synced: 6 months ago · JSON representation ·
Repository

Basic Info

Host: GitHub
Owner: R26raihan
Language: Jupyter Notebook
Default Branch: main
Size: 207 KB
Statistics

Stars: 0
Watchers: 0
Forks: 0
Open Issues: 0
Releases: 0
Created 7 months ago · Last pushed 7 months ago
Metadata Files

Readme Citation
Description

In this project, we addressed the task of identifying and classifying dataset citations in scientific articles as either Primary or Secondary. We began by loading the training labels and removing entries with missing dataset references. Then, we filtered the dataset to include only those articles with available XML files, from which we extracted the full article text by parsing
tags using Python’s XML utilities. After cleaning the data and ensuring proper text extraction, we performed exploratory data analysis to understand text lengths and label distribution.
We discovered a class imbalance, with Secondary citations significantly outnumbering Primary ones. To resolve this, we oversampled the underrepresented Primary class, creating a balanced dataset. A train-validation split was then performed to enable local model evaluation. Our next step involves training a transformer-based NLP model—such as SciBERT—fine-tuned on the labeled text data, to classify citation types based on the full article context. This setup prepares us to effectively tackle the citation classification challenge using state-of-the-art language modeling techniques.
Summary of the Modeling Workflow In this project, we developed a text classification model to identify whether a scientific article references a dataset as a Primary or Secondary citation. We began by exploring the labeled training data, where we found that the class distribution was imbalanced, with significantly more Secondary citations. To address this, we applied oversampling on the Primary class to create a balanced dataset. The text data from XML files was extracted and cleaned by lowercasing and removing punctuation. Then, we used TF-IDF vectorization with unigram and bigram features to convert the cleaned text into numerical form suitable for modeling.
After transforming the data, we trained and evaluated multiple machine learning models including Logistic Regression, Naive Bayes, Linear SVC, Random Forest, and XGBoost. The best-performing models achieved F1-scores above 0.93 on the validation set. Finally, we applied the trained model to the official unlabeled test set provided by the competition organizers, processed the XML files, and predicted the citation types for each article. The predictions were saved and are ready for submission or further analysis.
Owner

Login: R26raihan
Kind: user
Repositories: 1
Profile: https://github.com/R26raihan
Citation (citation-classification (1).ipynb)

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "028aea20",
   "metadata": {
    "papermill": {
     "duration": 0.005431,
     "end_time": "2025-07-05T12:25:38.003465",
     "exception": false,
     "start_time": "2025-07-05T12:25:37.998034",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Description \n",
    "\n",
    "In this project, we addressed the task of identifying and classifying dataset citations in scientific articles as either Primary or Secondary. We began by loading the training labels and removing entries with missing dataset references. Then, we filtered the dataset to include only those articles with available XML files, from which we extracted the full article text by parsing <p> tags using Python’s XML utilities. After cleaning the data and ensuring proper text extraction, we performed exploratory data analysis to understand text lengths and label distribution.\n",
    "\n",
    "We discovered a class imbalance, with Secondary citations significantly outnumbering Primary ones. To resolve this, we oversampled the underrepresented Primary class, creating a balanced dataset. A train-validation split was then performed to enable local model evaluation. Our next step involves training a transformer-based NLP model—such as SciBERT—fine-tuned on the labeled text data, to classify citation types based on the full article context. This setup prepares us to effectively tackle the citation classification challenge using state-of-the-art language modeling techniques."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a98b8d03",
   "metadata": {
    "papermill": {
     "duration": 0.004218,
     "end_time": "2025-07-05T12:25:38.012294",
     "exception": false,
     "start_time": "2025-07-05T12:25:38.008076",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Import Module"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "1405b945",
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5",
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:38.022119Z",
     "iopub.status.busy": "2025-07-05T12:25:38.021894Z",
     "iopub.status.idle": "2025-07-05T12:25:41.502477Z",
     "shell.execute_reply": "2025-07-05T12:25:41.501896Z"
    },
    "papermill": {
     "duration": 3.487091,
     "end_time": "2025-07-05T12:25:41.503800",
     "exception": false,
     "start_time": "2025-07-05T12:25:38.016709",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import os\n",
    "import glob\n",
    "import xml.etree.ElementTree as ET\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from bs4 import BeautifulSoup\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import re\n",
    "from sklearn.feature_extraction.text import TfidfVectorizer\n",
    "import joblib"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5c02b2ac",
   "metadata": {
    "papermill": {
     "duration": 0.004348,
     "end_time": "2025-07-05T12:25:41.512946",
     "exception": false,
     "start_time": "2025-07-05T12:25:41.508598",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Processing data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "3d98b437",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:41.522595Z",
     "iopub.status.busy": "2025-07-05T12:25:41.522266Z",
     "iopub.status.idle": "2025-07-05T12:25:41.546927Z",
     "shell.execute_reply": "2025-07-05T12:25:41.546234Z"
    },
    "papermill": {
     "duration": 0.030692,
     "end_time": "2025-07-05T12:25:41.547929",
     "exception": false,
     "start_time": "2025-07-05T12:25:41.517237",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total berlabel (bukan 'Missing'): 719\n"
     ]
    }
   ],
   "source": [
    "# ==== Step 1: Load Label ====\n",
    "labels_df = pd.read_csv('/kaggle/input/make-data-count-finding-data-references/train_labels.csv')\n",
    "\n",
    "# Ambil hanya yang tidak 'Missing'\n",
    "labels_df = labels_df[labels_df['dataset_id'] != 'Missing'].copy()\n",
    "print(\"Total berlabel (bukan 'Missing'):\", len(labels_df))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "411883f4",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:41.557583Z",
     "iopub.status.busy": "2025-07-05T12:25:41.557373Z",
     "iopub.status.idle": "2025-07-05T12:25:41.570707Z",
     "shell.execute_reply": "2025-07-05T12:25:41.569935Z"
    },
    "papermill": {
     "duration": 0.019283,
     "end_time": "2025-07-05T12:25:41.571748",
     "exception": false,
     "start_time": "2025-07-05T12:25:41.552465",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total file XML tersedia: 400\n"
     ]
    }
   ],
   "source": [
    "# ==== Step 2: Ambil semua file XML yang tersedia ====\n",
    "xml_dir = '/kaggle/input/make-data-count-finding-data-references/train/XML/'\n",
    "xml_files = glob.glob(os.path.join(xml_dir, '*.xml'))\n",
    "\n",
    "# Buat set nama file (tanpa .xml)\n",
    "available_article_ids = set(os.path.basename(f).replace('.xml', '') for f in xml_files)\n",
    "print(\"Total file XML tersedia:\", len(available_article_ids))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "46ff69b1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:41.581881Z",
     "iopub.status.busy": "2025-07-05T12:25:41.581622Z",
     "iopub.status.idle": "2025-07-05T12:25:41.592035Z",
     "shell.execute_reply": "2025-07-05T12:25:41.591407Z"
    },
    "papermill": {
     "duration": 0.016635,
     "end_time": "2025-07-05T12:25:41.593092",
     "exception": false,
     "start_time": "2025-07-05T12:25:41.576457",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Setelah cocok dengan file XML: 645\n"
     ]
    }
   ],
   "source": [
    "# ==== Step 3: Filter label yang ada file-nya ====\n",
    "labels_df = labels_df[labels_df['article_id'].isin(available_article_ids)].copy()\n",
    "print(\"Setelah cocok dengan file XML:\", len(labels_df))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "ccbaa3cf",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:41.603391Z",
     "iopub.status.busy": "2025-07-05T12:25:41.602793Z",
     "iopub.status.idle": "2025-07-05T12:25:41.608389Z",
     "shell.execute_reply": "2025-07-05T12:25:41.607820Z"
    },
    "papermill": {
     "duration": 0.011721,
     "end_time": "2025-07-05T12:25:41.609379",
     "exception": false,
     "start_time": "2025-07-05T12:25:41.597658",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# ==== Step 4: Normalisasi dataset_id ====\n",
    "def normalize_dataset_id(ds):\n",
    "    ds = str(ds).strip()\n",
    "    if ds.lower().startswith('doi:'):\n",
    "        ds = ds[4:]\n",
    "    if ds.lower().startswith('http://dx.doi.org/'):\n",
    "        ds = ds.replace('http://dx.doi.org/', '')\n",
    "    if ds.lower().startswith('doi.org/'):\n",
    "        ds = ds.replace('doi.org/', '')\n",
    "    if not ds.lower().startswith('https://doi.org/'):\n",
    "        return 'https://doi.org/' + ds\n",
    "    return ds\n",
    "\n",
    "labels_df['dataset_id'] = labels_df['dataset_id'].apply(normalize_dataset_id)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d690179b",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:41.619307Z",
     "iopub.status.busy": "2025-07-05T12:25:41.619122Z",
     "iopub.status.idle": "2025-07-05T12:25:41.624775Z",
     "shell.execute_reply": "2025-07-05T12:25:41.624087Z"
    },
    "papermill": {
     "duration": 0.011944,
     "end_time": "2025-07-05T12:25:41.626012",
     "exception": false,
     "start_time": "2025-07-05T12:25:41.614068",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# ==== Step 5: Fungsi ekstraksi teks lengkap dari XML (paragraf, tail, title, abstract) ====\n",
    "def extract_article_text(article_id, base_path=xml_dir):\n",
    "    file_path = os.path.join(base_path, f\"{article_id}.xml\")\n",
    "\n",
    "    try:\n",
    "        tree = ET.parse(file_path)\n",
    "        root = tree.getroot()\n",
    "        paragraphs = []\n",
    "\n",
    "        # Ambil teks dari <p> dan tail-nya\n",
    "        for elem in root.iter():\n",
    "            if elem.tag == 'p':\n",
    "                if elem.text:\n",
    "                    paragraphs.append(elem.text.strip())\n",
    "                for child in elem:\n",
    "                    if child.tail:\n",
    "                        paragraphs.append(child.tail.strip())\n",
    "\n",
    "        # Ambil judul\n",
    "        title = ''\n",
    "        title_elem = root.find(\".//title-group/article-title\")\n",
    "        if title_elem is not None and title_elem.text:\n",
    "            title = title_elem.text.strip()\n",
    "\n",
    "        # Ambil abstrak\n",
    "        abstract = ''\n",
    "        abstract_elem = root.find(\".//abstract\")\n",
    "        if abstract_elem is not None:\n",
    "            abstract = ''.join(abstract_elem.itertext()).strip()\n",
    "\n",
    "        full_text = \"\\n\".join([title, abstract] + paragraphs)\n",
    "        return full_text\n",
    "\n",
    "    except Exception as e:\n",
    "        print(f\"Gagal memproses {article_id}: {e}\")\n",
    "        return \"\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "e819f521",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:41.635973Z",
     "iopub.status.busy": "2025-07-05T12:25:41.635729Z",
     "iopub.status.idle": "2025-07-05T12:25:45.162227Z",
     "shell.execute_reply": "2025-07-05T12:25:45.161616Z"
    },
    "papermill": {
     "duration": 3.532893,
     "end_time": "2025-07-05T12:25:45.163481",
     "exception": false,
     "start_time": "2025-07-05T12:25:41.630588",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# ==== Step 6: Ambil teks semua artikel ====\n",
    "labels_df['text'] = labels_df['article_id'].apply(extract_article_text)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "cca3e10a",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:45.176521Z",
     "iopub.status.busy": "2025-07-05T12:25:45.176295Z",
     "iopub.status.idle": "2025-07-05T12:25:45.371083Z",
     "shell.execute_reply": "2025-07-05T12:25:45.370482Z"
    },
    "papermill": {
     "duration": 0.202552,
     "end_time": "2025-07-05T12:25:45.372447",
     "exception": false,
     "start_time": "2025-07-05T12:25:45.169895",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# ==== Step 7: (Opsional) Tandai apakah dataset_id muncul dalam teks ====\n",
    "labels_df['dataset_in_text'] = labels_df.apply(\n",
    "    lambda row: row['dataset_id'].lower() in row['text'].lower(), axis=1\n",
    ")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bf4ca97d",
   "metadata": {
    "papermill": {
     "duration": 0.00431,
     "end_time": "2025-07-05T12:25:45.381613",
     "exception": false,
     "start_time": "2025-07-05T12:25:45.377303",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Save Data labels"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "338fa3ce",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:45.391360Z",
     "iopub.status.busy": "2025-07-05T12:25:45.391144Z",
     "iopub.status.idle": "2025-07-05T12:25:45.992269Z",
     "shell.execute_reply": "2025-07-05T12:25:45.991474Z"
    },
    "papermill": {
     "duration": 0.607413,
     "end_time": "2025-07-05T12:25:45.993516",
     "exception": false,
     "start_time": "2025-07-05T12:25:45.386103",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "✅ Data berhasil disimpan ke processed_train_labeled.csv\n"
     ]
    }
   ],
   "source": [
    "# ==== Step 8: Simpan ke file CSV ====\n",
    "labels_df.to_csv(\"/kaggle/working/processed_train_labeled.csv\", index=False)\n",
    "print(\"✅ Data berhasil disimpan ke processed_train_labeled.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "71ce7b6e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:46.003756Z",
     "iopub.status.busy": "2025-07-05T12:25:46.003534Z",
     "iopub.status.idle": "2025-07-05T12:25:46.024717Z",
     "shell.execute_reply": "2025-07-05T12:25:46.024061Z"
    },
    "papermill": {
     "duration": 0.027472,
     "end_time": "2025-07-05T12:25:46.025815",
     "exception": false,
     "start_time": "2025-07-05T12:25:45.998343",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>article_id</th>\n",
       "      <th>dataset_id</th>\n",
       "      <th>type</th>\n",
       "      <th>dataset_in_text</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>10.1002_2017jc013030</td>\n",
       "      <td>https://doi.org/10.17882/49388</td>\n",
       "      <td>Primary</td>\n",
       "      <td>False</td>\n",
       "      <td>\\n</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>10.1002_ece3.4466</td>\n",
       "      <td>https://doi.org/10.5061/dryad.r6nq870</td>\n",
       "      <td>Primary</td>\n",
       "      <td>False</td>\n",
       "      <td>Incubation temperature and parental identity d...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>10.1002_ece3.5260</td>\n",
       "      <td>https://doi.org/10.5061/dryad.2f62927</td>\n",
       "      <td>Primary</td>\n",
       "      <td>False</td>\n",
       "      <td>Ultraconserved element (UCE) probe set design:...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>10.1002_ece3.6144</td>\n",
       "      <td>https://doi.org/10.5061/dryad.zw3r22854</td>\n",
       "      <td>Primary</td>\n",
       "      <td>False</td>\n",
       "      <td>Efficacy of metabarcoding for identification o...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>10.1002_ece3.6303</td>\n",
       "      <td>https://doi.org/10.5061/dryad.37pvmcvgb</td>\n",
       "      <td>Primary</td>\n",
       "      <td>False</td>\n",
       "      <td>Can polyploidy confer invasive plants with a w...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>10.1002_ece3.9627</td>\n",
       "      <td>https://doi.org/10.5061/dryad.b8gtht7h3</td>\n",
       "      <td>Primary</td>\n",
       "      <td>False</td>\n",
       "      <td>Inferring predator–prey interactions from came...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>10.1002_ecs2.4619</td>\n",
       "      <td>https://doi.org/10.25349/d9qw5x</td>\n",
       "      <td>Primary</td>\n",
       "      <td>False</td>\n",
       "      <td>\\n</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>10.1002_esp.5090</td>\n",
       "      <td>https://doi.org/10.5066/p9353101</td>\n",
       "      <td>Secondary</td>\n",
       "      <td>False</td>\n",
       "      <td>\\n</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>10.1002_mp.14424</td>\n",
       "      <td>https://doi.org/10.7937/k9/tcia.2015.pf0m9rei</td>\n",
       "      <td>Secondary</td>\n",
       "      <td>False</td>\n",
       "      <td>PleThora: Pleural effusion and thoracic cavity...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>10.1002_mp.14424</td>\n",
       "      <td>https://doi.org/10.7937/tcia.2020.6c7y-gq39</td>\n",
       "      <td>Primary</td>\n",
       "      <td>False</td>\n",
       "      <td>PleThora: Pleural effusion and thoracic cavity...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              article_id                                     dataset_id  \\\n",
       "0   10.1002_2017jc013030                 https://doi.org/10.17882/49388   \n",
       "11     10.1002_ece3.4466          https://doi.org/10.5061/dryad.r6nq870   \n",
       "12     10.1002_ece3.5260          https://doi.org/10.5061/dryad.2f62927   \n",
       "14     10.1002_ece3.6144        https://doi.org/10.5061/dryad.zw3r22854   \n",
       "15     10.1002_ece3.6303        https://doi.org/10.5061/dryad.37pvmcvgb   \n",
       "18     10.1002_ece3.9627        https://doi.org/10.5061/dryad.b8gtht7h3   \n",
       "20     10.1002_ecs2.4619                https://doi.org/10.25349/d9qw5x   \n",
       "25      10.1002_esp.5090               https://doi.org/10.5066/p9353101   \n",
       "26      10.1002_mp.14424  https://doi.org/10.7937/k9/tcia.2015.pf0m9rei   \n",
       "27      10.1002_mp.14424    https://doi.org/10.7937/tcia.2020.6c7y-gq39   \n",
       "\n",
       "         type  dataset_in_text  \\\n",
       "0     Primary            False   \n",
       "11    Primary            False   \n",
       "12    Primary            False   \n",
       "14    Primary            False   \n",
       "15    Primary            False   \n",
       "18    Primary            False   \n",
       "20    Primary            False   \n",
       "25  Secondary            False   \n",
       "26  Secondary            False   \n",
       "27    Primary            False   \n",
       "\n",
       "                                                 text  \n",
       "0                                                  \\n  \n",
       "11  Incubation temperature and parental identity d...  \n",
       "12  Ultraconserved element (UCE) probe set design:...  \n",
       "14  Efficacy of metabarcoding for identification o...  \n",
       "15  Can polyploidy confer invasive plants with a w...  \n",
       "18  Inferring predator–prey interactions from came...  \n",
       "20                                                 \\n  \n",
       "25                                                 \\n  \n",
       "26  PleThora: Pleural effusion and thoracic cavity...  \n",
       "27  PleThora: Pleural effusion and thoracic cavity...  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# ==== Step 9: Preview ====\n",
    "labels_df[['article_id', 'dataset_id', 'type', 'dataset_in_text', 'text']].head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "aad59304",
   "metadata": {
    "papermill": {
     "duration": 0.005354,
     "end_time": "2025-07-05T12:25:46.036149",
     "exception": false,
     "start_time": "2025-07-05T12:25:46.030795",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Explatory data Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "1fdd6d76",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:46.046338Z",
     "iopub.status.busy": "2025-07-05T12:25:46.046131Z",
     "iopub.status.idle": "2025-07-05T12:25:46.668093Z",
     "shell.execute_reply": "2025-07-05T12:25:46.667327Z"
    },
    "papermill": {
     "duration": 0.628296,
     "end_time": "2025-07-05T12:25:46.669239",
     "exception": false,
     "start_time": "2025-07-05T12:25:46.040943",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(data=labels_df, x='type')\n",
    "plt.title(\"Distribusi Label\")\n",
    "plt.show()\n",
    "\n",
    "# Panjang teks artikel\n",
    "labels_df['text_length'] = labels_df['text'].apply(lambda x: len(x.split()))\n",
    "labels_df['text_length'].hist(bins=50)\n",
    "plt.title(\"Distribusi Panjang Teks Artikel\")\n",
    "plt.xlabel(\"Jumlah Kata\")\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "218cf14c",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:46.682353Z",
     "iopub.status.busy": "2025-07-05T12:25:46.682139Z",
     "iopub.status.idle": "2025-07-05T12:25:46.862774Z",
     "shell.execute_reply": "2025-07-05T12:25:46.861976Z"
    },
    "papermill": {
     "duration": 0.188384,
     "end_time": "2025-07-05T12:25:46.863934",
     "exception": false,
     "start_time": "2025-07-05T12:25:46.675550",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Jumlah artikel dengan teks kosong: 148\n"
     ]
    }
   ],
   "source": [
    "labels_df['text_length'] = labels_df['text'].apply(lambda x: len(x.strip().split()))\n",
    "print(\"Jumlah artikel dengan teks kosong:\", (labels_df['text_length'] == 0).sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "962415c0",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:46.876516Z",
     "iopub.status.busy": "2025-07-05T12:25:46.876305Z",
     "iopub.status.idle": "2025-07-05T12:25:46.880642Z",
     "shell.execute_reply": "2025-07-05T12:25:46.880158Z"
    },
    "papermill": {
     "duration": 0.012035,
     "end_time": "2025-07-05T12:25:46.881679",
     "exception": false,
     "start_time": "2025-07-05T12:25:46.869644",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "\n",
    "def extract_article_text_bs(article_id, base_path=xml_dir):\n",
    "    file_path = os.path.join(base_path, f\"{article_id}.xml\")\n",
    "    try:\n",
    "        with open(file_path, 'r', encoding='utf-8') as f:\n",
    "            xml = f.read()\n",
    "            soup = BeautifulSoup(xml, 'xml')\n",
    "            # Ambil semua teks dari <p> tag\n",
    "            paragraphs = [p.get_text(strip=True) for p in soup.find_all('p')]\n",
    "            return \"\\n\".join(paragraphs)\n",
    "    except Exception as e:\n",
    "        print(f\"Gagal memproses {article_id}: {e}\")\n",
    "        return \"\"\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "731c6189",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:25:46.893438Z",
     "iopub.status.busy": "2025-07-05T12:25:46.893034Z",
     "iopub.status.idle": "2025-07-05T12:26:31.056120Z",
     "shell.execute_reply": "2025-07-05T12:26:31.055254Z"
    },
    "papermill": {
     "duration": 44.175776,
     "end_time": "2025-07-05T12:26:31.062953",
     "exception": false,
     "start_time": "2025-07-05T12:25:46.887177",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Jumlah artikel dengan teks kosong setelah parsing ulang: 24\n"
     ]
    }
   ],
   "source": [
    "labels_df['text'] = labels_df['article_id'].apply(extract_article_text_bs)\n",
    "labels_df['text_length'] = labels_df['text'].apply(lambda x: len(x.strip().split()))\n",
    "print(\"Jumlah artikel dengan teks kosong setelah parsing ulang:\", (labels_df['text_length'] == 0).sum())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "70e4e7b3",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:31.075917Z",
     "iopub.status.busy": "2025-07-05T12:26:31.075649Z",
     "iopub.status.idle": "2025-07-05T12:26:31.233937Z",
     "shell.execute_reply": "2025-07-05T12:26:31.233258Z"
    },
    "papermill": {
     "duration": 0.16641,
     "end_time": "2025-07-05T12:26:31.235151",
     "exception": false,
     "start_time": "2025-07-05T12:26:31.068741",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Axes: title={'center': 'Distribusi Label (Primary vs Secondary)'}, xlabel='type'>"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels_df['type'].value_counts().plot(kind='bar', title=\"Distribusi Label (Primary vs Secondary)\", rot=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "83cb0180",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:31.248488Z",
     "iopub.status.busy": "2025-07-05T12:26:31.248295Z",
     "iopub.status.idle": "2025-07-05T12:26:31.658662Z",
     "shell.execute_reply": "2025-07-05T12:26:31.658015Z"
    },
    "papermill": {
     "duration": 0.418252,
     "end_time": "2025-07-05T12:26:31.659959",
     "exception": false,
     "start_time": "2025-07-05T12:26:31.241707",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "labels_df['word_count'] = labels_df['text'].apply(lambda x: len(x.strip().split()))\n",
    "labels_df['word_count'].hist(bins=50, figsize=(10, 5))\n",
    "plt.title(\"Distribusi Jumlah Kata\")\n",
    "plt.xlabel(\"Jumlah Kata\")\n",
    "plt.ylabel(\"Jumlah Artikel\")\n",
    "plt.grid(True)\n",
    "plt.show()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "0e6e320e",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:31.674216Z",
     "iopub.status.busy": "2025-07-05T12:26:31.673635Z",
     "iopub.status.idle": "2025-07-05T12:26:31.681469Z",
     "shell.execute_reply": "2025-07-05T12:26:31.680704Z"
    },
    "papermill": {
     "duration": 0.015628,
     "end_time": "2025-07-05T12:26:31.682586",
     "exception": false,
     "start_time": "2025-07-05T12:26:31.666958",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Contoh teks pendek:\n",
      "\n",
      "\n",
      "Contoh teks panjang:\n",
      "These authors contributed equally to this work.\n",
      "Bioinformatics and Data Science, Research and Early Development, Oncology R&D, AstraZeneca, Cambridge, United Kingdom.\n",
      "Data Science & Artificial Intelligence, Imaging & Data Analytics, Clinical Pharmacology & Safety Sciences, R&D, AstraZeneca, Gothenburg, Sweden.\n",
      "Prostate cancer is the second most occurring cancer in men worldwide. To better understand the mechanisms of tumorigenesis and possible treatment responses, we developed a mathematical model of prostate cancer which considers the major signalling pathways known to be deregulated. We personalised this Boolean model to molecular data to reflect the heterogeneity and specific response to perturbations of cancer patients. A total of 488 prostate samples were used to build patient-specific models and compared to available clinical data. Additionally, eight prostate cell line-specific models were built to validate our approach with dose-response data of several drugs. The effects of si\n"
     ]
    }
   ],
   "source": [
    "print(\"Contoh teks pendek:\")\n",
    "print(labels_df.sort_values('word_count').iloc[0]['text'])\n",
    "\n",
    "print(\"\\nContoh teks panjang:\")\n",
    "print(labels_df.sort_values('word_count', ascending=False).iloc[0]['text'][:1000])  # max 1000 char\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "70f847e7",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:31.696845Z",
     "iopub.status.busy": "2025-07-05T12:26:31.696666Z",
     "iopub.status.idle": "2025-07-05T12:26:31.704467Z",
     "shell.execute_reply": "2025-07-05T12:26:31.703924Z"
    },
    "papermill": {
     "duration": 0.015913,
     "end_time": "2025-07-05T12:26:31.705543",
     "exception": false,
     "start_time": "2025-07-05T12:26:31.689630",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count      645.000000\n",
       "mean      6816.348837\n",
       "std       4451.802377\n",
       "min          0.000000\n",
       "25%       4040.000000\n",
       "50%       6138.000000\n",
       "75%       7765.000000\n",
       "max      27188.000000\n",
       "Name: word_count, dtype: float64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "labels_df['word_count'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "83f76862",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:31.719648Z",
     "iopub.status.busy": "2025-07-05T12:26:31.719435Z",
     "iopub.status.idle": "2025-07-05T12:26:31.857829Z",
     "shell.execute_reply": "2025-07-05T12:26:31.857082Z"
    },
    "papermill": {
     "duration": 0.146729,
     "end_time": "2025-07-05T12:26:31.859186",
     "exception": false,
     "start_time": "2025-07-05T12:26:31.712457",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "labels_df['type'].value_counts().plot(kind='bar', title=\"Distribusi Label Citation Type\")\n",
    "plt.xlabel(\"Citation Type\")\n",
    "plt.ylabel(\"Jumlah\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "25677faf",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:31.874974Z",
     "iopub.status.busy": "2025-07-05T12:26:31.874425Z",
     "iopub.status.idle": "2025-07-05T12:26:31.884555Z",
     "shell.execute_reply": "2025-07-05T12:26:31.883798Z"
    },
    "papermill": {
     "duration": 0.019123,
     "end_time": "2025-07-05T12:26:31.885676",
     "exception": false,
     "start_time": "2025-07-05T12:26:31.866553",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "# Misal: class 'Primary' underrepresented\n",
    "primary_df = labels_df[labels_df['type'] == 'Primary']\n",
    "secondary_df = labels_df[labels_df['type'] == 'Secondary']\n",
    "\n",
    "# Oversample Primary secara manual\n",
    "primary_oversampled = primary_df.sample(len(secondary_df), replace=True, random_state=42)\n",
    "\n",
    "# Gabung jadi dataset seimbang\n",
    "balanced_df = pd.concat([secondary_df, primary_oversampled]).sample(frac=1, random_state=42).reset_index(drop=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "88fca42d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:31.901182Z",
     "iopub.status.busy": "2025-07-05T12:26:31.900955Z",
     "iopub.status.idle": "2025-07-05T12:26:32.041013Z",
     "shell.execute_reply": "2025-07-05T12:26:32.040224Z"
    },
    "papermill": {
     "duration": 0.149114,
     "end_time": "2025-07-05T12:26:32.042223",
     "exception": false,
     "start_time": "2025-07-05T12:26:31.893109",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "balanced_df['type'].value_counts().plot(kind='bar', title=\"Distribusi Label Citation Type\")\n",
    "plt.xlabel(\"Citation Type\")\n",
    "plt.ylabel(\"Jumlah\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c8814bbf",
   "metadata": {
    "papermill": {
     "duration": 0.007623,
     "end_time": "2025-07-05T12:26:32.058268",
     "exception": false,
     "start_time": "2025-07-05T12:26:32.050645",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "During the exploratory data analysis (EDA) stage, we identified several key insights related to data quality and distribution. Initially, we discovered that 148 articles had empty text content after XML parsing. Upon reprocessing using a more robust parsing method, this number was reduced to 24. This highlighted inconsistencies in the XML structure across articles, emphasizing the need for flexible parsing techniques to ensure text extraction is as complete and accurate as possible.\n",
    "\n",
    "Additionally, we observed a significant class imbalance in the citation type labels — the Secondary class appeared far more frequently than Primary. This imbalance posed a risk of model bias toward the majority class during training. To address this, we applied oversampling to the Primary class to equalize the number of samples from both classes. Balancing the dataset ensures that the classifier can learn representative features for both citation types, improving overall performance and fairness."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "610f28b1",
   "metadata": {
    "papermill": {
     "duration": 0.007411,
     "end_time": "2025-07-05T12:26:32.073330",
     "exception": false,
     "start_time": "2025-07-05T12:26:32.065919",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "#  Preprocessing Teks and Vectorizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "37e9ee5d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:32.089447Z",
     "iopub.status.busy": "2025-07-05T12:26:32.089037Z",
     "iopub.status.idle": "2025-07-05T12:26:44.034454Z",
     "shell.execute_reply": "2025-07-05T12:26:44.033822Z"
    },
    "papermill": {
     "duration": 11.954855,
     "end_time": "2025-07-05T12:26:44.035681",
     "exception": false,
     "start_time": "2025-07-05T12:26:32.080826",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['tfidf_vectorizer.pkl']"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "def clean_text(text):\n",
    "    text = text.lower()\n",
    "    text = re.sub(r'https?://\\S+|www\\.\\S+', '', text)\n",
    "    text = re.sub(r'[^a-z\\s]', '', text)\n",
    "    text = re.sub(r'\\s+', ' ', text).strip()\n",
    "    return text\n",
    "\n",
    "balanced_df['clean_text'] = balanced_df['text'].fillna('').apply(clean_text)\n",
    "\n",
    "tfidf = TfidfVectorizer(max_features=10000, ngram_range=(1, 2))\n",
    "X = tfidf.fit_transform(balanced_df['clean_text'])\n",
    "y = balanced_df['type']\n",
    "\n",
    "# Simpan vectorizer\n",
    "joblib.dump(tfidf, 'tfidf_vectorizer.pkl')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "e929d4b1",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:44.053101Z",
     "iopub.status.busy": "2025-07-05T12:26:44.052826Z",
     "iopub.status.idle": "2025-07-05T12:26:44.159382Z",
     "shell.execute_reply": "2025-07-05T12:26:44.158504Z"
    },
    "papermill": {
     "duration": 0.116278,
     "end_time": "2025-07-05T12:26:44.160599",
     "exception": false,
     "start_time": "2025-07-05T12:26:44.044321",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "TF-IDF matrix shape: (846, 10000)\n",
      "\n",
      "Contoh representasi TF-IDF (5 dokumen pertama):\n",
      "[[0.         0.00304524 0.         ... 0.         0.         0.        ]\n",
      " [0.         0.         0.         ... 0.         0.         0.        ]\n",
      " [0.         0.         0.         ... 0.         0.         0.        ]\n",
      " [0.00232796 0.         0.         ... 0.         0.         0.        ]\n",
      " [0.         0.         0.         ... 0.         0.         0.        ]]\n",
      "\n",
      "Fitur (kata/ngram) yang dipakai:\n",
      "['aa' 'ab' 'aba' 'abbreviated' 'abbreviated as' 'abc' 'ability'\n",
      " 'ability of' 'ability to' 'abiotic' 'able' 'able to' 'abnormal' 'about'\n",
      " 'about the' 'above' 'above and' 'above the' 'above we' 'aboveground']\n"
     ]
    }
   ],
   "source": [
    "# 1. Bentuk matriks hasil vektorisasi\n",
    "print(\"TF-IDF matrix shape:\", X.shape)  # (jumlah dokumen, jumlah fitur)\n",
    "\n",
    "# 2. Konversi ke array agar bisa dilihat nilai-nilainya\n",
    "X_dense = X.todense()\n",
    "\n",
    "# 3. Lihat beberapa baris pertama\n",
    "print(\"\\nContoh representasi TF-IDF (5 dokumen pertama):\")\n",
    "print(X_dense[:5])\n",
    "\n",
    "# 4. Tampilkan beberapa fitur kata yang digunakan\n",
    "print(\"\\nFitur (kata/ngram) yang dipakai:\")\n",
    "print(tfidf.get_feature_names_out()[:20])  # 20 fitur pertama\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a99fd68",
   "metadata": {
    "papermill": {
     "duration": 0.007855,
     "end_time": "2025-07-05T12:26:44.176776",
     "exception": false,
     "start_time": "2025-07-05T12:26:44.168921",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Modelling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "6bf3e08d",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:44.193182Z",
     "iopub.status.busy": "2025-07-05T12:26:44.192946Z",
     "iopub.status.idle": "2025-07-05T12:26:53.786356Z",
     "shell.execute_reply": "2025-07-05T12:26:53.785438Z"
    },
    "papermill": {
     "duration": 9.603943,
     "end_time": "2025-07-05T12:26:53.788489",
     "exception": false,
     "start_time": "2025-07-05T12:26:44.184546",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "🔍 Model: Logistic Regression\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "     Primary       0.91      0.95      0.93        85\n",
      "   Secondary       0.95      0.91      0.93        85\n",
      "\n",
      "    accuracy                           0.93       170\n",
      "   macro avg       0.93      0.93      0.93       170\n",
      "weighted avg       0.93      0.93      0.93       170\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "🔍 Model: Naive Bayes\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "     Primary       0.87      0.93      0.90        85\n",
      "   Secondary       0.92      0.86      0.89        85\n",
      "\n",
      "    accuracy                           0.89       170\n",
      "   macro avg       0.90      0.89      0.89       170\n",
      "weighted avg       0.90      0.89      0.89       170\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "🔍 Model: Linear SVC\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "     Primary       0.96      0.94      0.95        85\n",
      "   Secondary       0.94      0.96      0.95        85\n",
      "\n",
      "    accuracy                           0.95       170\n",
      "   macro avg       0.95      0.95      0.95       170\n",
      "weighted avg       0.95      0.95      0.95       170\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "🔍 Model: Random Forest\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "     Primary       0.95      0.94      0.95        85\n",
      "   Secondary       0.94      0.95      0.95        85\n",
      "\n",
      "    accuracy                           0.95       170\n",
      "   macro avg       0.95      0.95      0.95       170\n",
      "weighted avg       0.95      0.95      0.95       170\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "🔍 Model: XGBoost\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "     Primary       0.96      0.94      0.95        85\n",
      "   Secondary       0.94      0.96      0.95        85\n",
      "\n",
      "    accuracy                           0.95       170\n",
      "   macro avg       0.95      0.95      0.95       170\n",
      "weighted avg       0.95      0.95      0.95       170\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "💾 Menyimpan model terbaik: Linear SVC (Accuracy: 0.9529)\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.linear_model import LogisticRegression\n",
    "from sklearn.naive_bayes import MultinomialNB\n",
    "from sklearn.svm import LinearSVC\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import classification_report, confusion_matrix, ConfusionMatrixDisplay\n",
    "from sklearn.model_selection import train_test_split\n",
    "from xgboost import XGBClassifier\n",
    "import matplotlib.pyplot as plt\n",
    "import joblib  # ✅ untuk menyimpan model dan vectorizer\n",
    "\n",
    "# Encode y menjadi angka\n",
    "le = LabelEncoder()\n",
    "y_encoded = le.fit_transform(y)  # 'Primary' -> 0, 'Secondary' -> 1\n",
    "\n",
    "# Split data\n",
    "X_train, X_val, y_train, y_val = train_test_split(X, y_encoded, stratify=y_encoded, test_size=0.2, random_state=42)\n",
    "\n",
    "# Daftar model\n",
    "models = {\n",
    "    \"Logistic Regression\": LogisticRegression(max_iter=1000),\n",
    "    \"Naive Bayes\": MultinomialNB(),\n",
    "    \"Linear SVC\": LinearSVC(),\n",
    "    \"Random Forest\": RandomForestClassifier(n_estimators=100, random_state=42),\n",
    "    \"XGBoost\": XGBClassifier(use_label_encoder=False, eval_metric='logloss', random_state=42)\n",
    "}\n",
    "\n",
    "best_model = None\n",
    "best_score = 0\n",
    "\n",
    "# Evaluasi model\n",
    "for name, model in models.items():\n",
    "    print(f\"\\n🔍 Model: {name}\")\n",
    "    model.fit(X_train, y_train)\n",
    "    y_pred = model.predict(X_val)\n",
    "\n",
    "    # Classification report\n",
    "    print(classification_report(le.inverse_transform(y_val), le.inverse_transform(y_pred)))\n",
    "\n",
    "    # Confusion matrix\n",
    "    cm = confusion_matrix(y_val, y_pred)\n",
    "    disp = ConfusionMatrixDisplay(confusion_matrix=cm, display_labels=le.classes_)\n",
    "    disp.plot(cmap='Blues')\n",
    "    plt.title(f'Confusion Matrix - {name}')\n",
    "    plt.show()\n",
    "\n",
    "    # Simpan model terbaik berdasarkan akurasi\n",
    "    score = model.score(X_val, y_val)\n",
    "    if score > best_score:\n",
    "        best_score = score\n",
    "        best_model = model\n",
    "        best_model_name = name\n",
    "\n",
    "# ✅ Simpan model terbaik, tfidf vectorizer, dan label encoder\n",
    "if best_model is not None:\n",
    "    print(f\"\\n💾 Menyimpan model terbaik: {best_model_name} (Accuracy: {best_score:.4f})\")\n",
    "    joblib.dump(best_model, 'best_model.pkl')\n",
    "    joblib.dump(tfidf, 'tfidf_vectorizer.pkl')\n",
    "    joblib.dump(le, 'label_encoder.pkl')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "83674b8e",
   "metadata": {
    "papermill": {
     "duration": 0.010733,
     "end_time": "2025-07-05T12:26:53.812708",
     "exception": false,
     "start_time": "2025-07-05T12:26:53.801975",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "# Predicting on Test Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "11a843aa",
   "metadata": {
    "execution": {
     "iopub.execute_input": "2025-07-05T12:26:53.835816Z",
     "iopub.status.busy": "2025-07-05T12:26:53.835564Z",
     "iopub.status.idle": "2025-07-05T12:26:54.730790Z",
     "shell.execute_reply": "2025-07-05T12:26:54.730218Z"
    },
    "papermill": {
     "duration": 0.908626,
     "end_time": "2025-07-05T12:26:54.732207",
     "exception": false,
     "start_time": "2025-07-05T12:26:53.823581",
     "status": "completed"
    },
    "tags": []
   },
   "outputs": [],
   "source": [
    "import os\n",
    "import glob\n",
    "import xml.etree.ElementTree as ET\n",
    "import pandas as pd\n",
    "import re\n",
    "import joblib\n",
    "\n",
    "# Load kembali vectorizer, model, dan encoder\n",
    "tfidf = joblib.load('/kaggle/working/tfidf_vectorizer.pkl')\n",
    "model = joblib.load('/kaggle/working/best_model.pkl')\n",
    "le = joblib.load('/kaggle/working/label_encoder.pkl')\n",
    "\n",
    "# Ambil semua XML test\n",
    "test_dir = '/kaggle/input/make-data-count-finding-data-references/test/XML/'\n",
    "test_files = glob.glob(os.path.join(test_dir, '*.xml'))\n",
    "test_ids = [os.path.basename(f).replace('.xml', '') for f in test_files]\n",
    "\n",
    "def extract_article_text(article_id, base_path=test_dir):\n",
    "    file_path = os.path.join(base_path, f\"{article_id}.xml\")\n",
    "    try:\n",
    "        tree = ET.parse(file_path)\n",
    "        root = tree.getroot()\n",
    "        paragraphs = [elem.text.strip() for elem in root.iter() if elem.tag == 'p' and elem.text]\n",
    "        return \"\\n\".join(paragraphs)\n",
    "    except:\n",
    "        return \"\"\n",
    "\n",
    "def clean_text(text):\n",
    "    text = text.lower()\n",
    "    text = re.sub(r'\\s+', ' ', text)\n",
    "    text = re.sub(r'[^a-zA-Z0-9\\s]', '', text)\n",
    "    return text.strip()\n",
    "\n",
    "# Proses test data\n",
    "test_df = pd.DataFrame({'article_id': test_ids})\n",
    "test_df['text'] = test_df['article_id'].apply(extract_article_text)\n",
    "test_df['clean_text'] = test_df['text'].apply(clean_text)\n",
    "\n",
    "# Transform dan prediksi\n",
    "X_test = tfidf.transform(test_df['clean_text'])\n",
    "y_test_pred = model.predict(X_test)\n",
    "\n",
    "# Konversi label\n",
    "predicted_labels = le.inverse_transform(y_test_pred)\n",
    "test_df['predicted_type'] = predicted_labels\n",
    "\n",
    "# Simpan sebagai submission\n",
    "test_df[['article_id', 'predicted_type']].to_csv('/kaggle/working/submission.csv', index=False)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48085029",
   "metadata": {
    "papermill": {
     "duration": 0.011487,
     "end_time": "2025-07-05T12:26:54.757047",
     "exception": false,
     "start_time": "2025-07-05T12:26:54.745560",
     "status": "completed"
    },
    "tags": []
   },
   "source": [
    "Summary of the Modeling Workflow\n",
    "In this project, we developed a text classification model to identify whether a scientific article references a dataset as a Primary or Secondary citation. We began by exploring the labeled training data, where we found that the class distribution was imbalanced, with significantly more Secondary citations. To address this, we applied oversampling on the Primary class to create a balanced dataset. The text data from XML files was extracted and cleaned by lowercasing and removing punctuation. Then, we used TF-IDF vectorization with unigram and bigram features to convert the cleaned text into numerical form suitable for modeling.\n",
    "\n",
    "After transforming the data, we trained and evaluated multiple machine learning models including Logistic Regression, Naive Bayes, Linear SVC, Random Forest, and XGBoost. The best-performing models achieved F1-scores above 0.93 on the validation set. Finally, we applied the trained model to the official unlabeled test set provided by the competition organizers, processed the XML files, and predicted the citation types for each article. The predictions were saved and are ready for submission or further analysis."
   ]
  }
 ],
 "metadata": {
  "kaggle": {
   "accelerator": "nvidiaTeslaT4",
   "dataSources": [
    {
     "databundleVersionId": 12656064,
     "sourceId": 82370,
     "sourceType": "competition"
    }
   ],
   "dockerImageVersionId": 31040,
   "isGpuEnabled": true,
   "isInternetEnabled": false,
   "language": "python",
   "sourceType": "notebook"
  },
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.11.11"
  },
  "papermill": {
   "default_parameters": {},
   "duration": 81.679132,
   "end_time": "2025-07-05T12:26:55.486233",
   "environment_variables": {},
   "exception": null,
   "input_path": "__notebook__.ipynb",
   "output_path": "__notebook__.ipynb",
   "parameters": {},
   "start_time": "2025-07-05T12:25:33.807101",
   "version": "2.6.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}
ecosyste.ms

Data

Tools

Indexes

Applications

Experiments

Open Source Science

dml

Science Score: 44.0%

Repository

Basic Info

Statistics

Metadata Files

README.md

Description

Owner

Citation (citation-classification (1).ipynb)

GitHub Events

Total

Last Year
