{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "a5b9ff56-9501-464c-8be0-c9ca203d0f5e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Age average:  30\n"
     ]
    }
   ],
   "source": [
    "#Bagging\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "from sklearn.ensemble import BaggingClassifier\n",
    "from sklearn.model_selection import train_test_split\n",
    "import sklearn.metrics as metrics\n",
    "\n",
    "# Import data of the titanic passengers\n",
    "data_titanic = pd.read_csv(\"titanic_data.csv\")\n",
    "\n",
    "# Preprocessing\n",
    "# Not all data is important for the training. The PassengerId, the name, the number of siblings, the parch, the ticket number\n",
    "# and the cabin number are not important and therefore they will be dropped\n",
    "data_titanic = data_titanic.drop(data_titanic.columns[[0, 3, 6, 7, 8, 10]], axis = 1)\n",
    "\n",
    "# Fill the gaps:\n",
    "# For the age: Compute the average of ages of all passengers on board and take it as value\n",
    "ages = data_titanic[\"Age\"]\n",
    "age_average = round(ages.mean(axis = 0,  skipna = True)) \n",
    "print('Age average: ', age_average)\n",
    "\n",
    "def set_age(Age):\n",
    "    age = Age\n",
    "    if pd.isnull(age):\n",
    "        return age_average\n",
    "    else:\n",
    "        return age\n",
    "\n",
    "data_titanic['Age'] = data_titanic['Age'].apply(set_age)\n",
    "    \n",
    "# For the embarked: Take the most occuring value, available options: S, C and Q\n",
    "Embarked = data_titanic[\"Embarked\"]\n",
    "count_S = 0\n",
    "count_C = 0\n",
    "count_Q = 0\n",
    "\n",
    "# Count embarks\n",
    "for i in range(data_titanic.shape[0]):\n",
    "    if (Embarked[i] == 'S'):\n",
    "        count_S += 1\n",
    "    if (Embarked[i] == 'C'):\n",
    "        count_C += 1\n",
    "    if (Embarked[i] == 'Q'):\n",
    "        count_Q += 1\n",
    "        \n",
    "if ((count_S >= count_C) and (count_S >= count_Q)):\n",
    "    common_embarked = 'S'\n",
    "if ((count_C > count_S) and (count_C >= count_Q)):\n",
    "    common_embarked = 'C'\n",
    "if ((count_Q >= count_S) and (count_Q >= count_C)):\n",
    "    common_embarked = 'Q'\n",
    "        \n",
    "def set_Embarked(Embarked):\n",
    "    embarked = Embarked\n",
    "    if pd.isnull(embarked):\n",
    "        return common_embarked\n",
    "    else:\n",
    "        return embarked\n",
    "\n",
    "data_titanic['Embarked'] = data_titanic['Embarked'].apply(set_Embarked)\n",
    "\n",
    "# Gender and Embarked have to be replaced by values: Set male = 0, female = 1 and S = 0, C = 1, Q = 2\n",
    "replace_gender = {data_titanic.columns[2]:     {\"male\": 0, \"female\": 1}}\n",
    "replace_embarked = {data_titanic.columns[5]:     {\"S\": 0, \"C\": 1, \"Q\": 2}}\n",
    "data_titanic = data_titanic.replace(replace_gender)\n",
    "data_titanic = data_titanic.replace(replace_embarked)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "2a63a9f4-b658-4556-8cb2-6fa3eba5ce98",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Split data into train and evaluation subsets, proportion 1:4\n",
    "input_data = data_titanic.drop(data_titanic.columns[[0]], axis = 1) # Input has to be without \"Survived\" statement\n",
    "target_data = data_titanic['Survived'] # KNN algorithm has to predict survied statement\n",
    "input_train, input_evaluate, target_train, target_evaluate = train_test_split(input_data , target_data, test_size = 0.25)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "c3575857-7cb9-4a23-a9b3-89ae47e896ba",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy:  0.8071748878923767\n"
     ]
    }
   ],
   "source": [
    "# Perform Bagging\n",
    "classifier = BaggingClassifier(n_estimators = 20, random_state=0)\n",
    "classifier.fit(input_train,target_train)\n",
    "target_pred = classifier.predict(input_evaluate)\n",
    "print(\"Accuracy: \", metrics.accuracy_score(target_pred, target_evaluate))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "767d178f-60cc-41d4-a698-9c47344fcada",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.10.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}