Merge branch 'main' of https://gitty.informatik.hs-mannheim.de/1826514/DSA_SS24

README.md

@@ -218,11 +218,12 @@ The selection of features was informed by an analysis presented in a paper (sour
 The exact process can be found in the notebook: [features_detection.ipynb](notebooks/features_detection.ipynb).
 
 ### ML-models
-For machine learning, the initial step involved tailoring the features for the models, followed by employing a grid search to identify the best hyperparameters. This approach led to the highest performance being achieved by the Extreme Gradient Boosting (XGBoost) model, which attained an accuracy of 83%. Additionally, a Gradient Boosting Tree model was evaluated using the same procedure and achieved an accuracy of 82%. The selection of these models was influenced by the team's own experience and the performance metrics highlighted in the paper (source: https://rdcu.be/dH2jI, last accessed: 15.05.2024). The models have also been evaluated, and it is noticeable that some features, like the ventricular rate, are shown to be more important than other features.
+For machine learning, the initial step involved tailoring the features for the models, followed by employing a grid search to identify the best hyperparameters. This approach led to the highest performance being achieved by the Extreme Gradient Boosting (XGBoost) model, which attained an accuracy of 83%. Additionally, a Gradient Boosting Tree model was evaluated using the same procedure and achieved an accuracy of 82%. A Decision Tree model was also evaluated, having the lowest performance of 80%. The selection of these models was influenced by the team's own experience and the performance metrics highlighted in the paper (source: https://rdcu.be/dH2jI, last accessed: 15.05.2024). The models have also been evaluated, and it is noticeable that some features, like the ventricular rate, are shown to be more important than other features.
 
 <br>The detailed procedures can be found in the following notebooks: 
 <br>[ml_xgboost.ipynb](notebooks/ml_xgboost.ipynb)
 <br>[ml_grad_boost_tree.ipynb](notebooks/ml_grad_boost_tree.ipynb)
+<br>[ml_decision_tree.ipynb](notebooks/ml_decision_tree.ipynb)
 
 
 ## Contributing
@@ -244,6 +245,15 @@ Please note, by contributing to this project, you agree that your contributions
 We look forward to your contributions. Thank you for helping us improve this project!
 
 
+# Update from 03.07
+
+## Conclusion
+- Machine learning and data analysis as valuable tools for investigating cardiovascular diseases 
+
+ - Improvement of diagnostics and treatment possible through predictive modeling
+
+## Outlook into the future
+
 ## License
 This project is licensed under the [MIT License](https://opensource.org/licenses/MIT).
 

notebooks/decision_tree.ipynb

@@ -0,0 +1,207 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Decison Tree"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import sqlite3\n",
+    "import os\n",
+    "from datetime import datetime\n",
+    "from joblib import dump, load\n",
+    "import pandas as pd\n",
+    "import matplotlib.pyplot as plt\n",
+    "import xgboost as xgb\n",
+    "from sklearn.model_selection import GridSearchCV\n",
+    "from sklearn.metrics import confusion_matrix, f1_score\n",
+    "from sklearn.ensemble import GradientBoostingClassifier\n",
+    "from sklearn.impute import SimpleImputer\n",
+    "from sklearn.metrics import accuracy_score\n",
+    "from sklearn.preprocessing import MinMaxScaler\n",
+    "\n",
+    "import seaborn as sns"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Import Data from Database"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# connect to the database\n",
+    "conn = sqlite3.connect('../features.db')\n",
+    "c = conn.cursor()\n",
+    "# get training, validation and test data\n",
+    "train = pd.read_sql_query(\"SELECT * FROM train\", conn)\n",
+    "valid = pd.read_sql_query(\"SELECT * FROM validation\", conn)\n",
+    "test = pd.read_sql_query(\"SELECT * FROM test\", conn)\n",
+    "# close the connection\n",
+    "conn.close()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Format Data for Machine Learning"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "train_x shape: (3502, 10)\n",
+      "test_x shape: (438, 10)\n",
+      "valid_x shape: (438, 10)\n",
+      "features: ['age', 'gender', 'artial_rate', 'ventricular_rate', 'qrs_duration', 'qt_length', 'qrs_count', 'q_peak', 'r_axis', 't_axis']\n",
+      "number of classes: 4\n"
+     ]
+    }
+   ],
+   "source": [
+    "# get the target and features\n",
+    "train_y = train['y']\n",
+    "train_y = train_y.map({'GSVT': 0, 'AFIB': 1, 'SR': 2, 'SB': 3})\n",
+    "train_x = train.drop(columns=['y'])\n",
+    "\n",
+    "valid_y = valid['y']\n",
+    "valid_y = valid_y.map({'GSVT': 0, 'AFIB': 1, 'SR': 2, 'SB': 3})\n",
+    "valid_x = valid.drop(columns=['y'])\n",
+    "\n",
+    "test_y = test['y']\n",
+    "test_y = test_y.map({'GSVT': 0, 'AFIB': 1, 'SR': 2, 'SB': 3})\n",
+    "test_x = test.drop(columns=['y'])\n",
+    "\n",
+    "# drop id column\n",
+    "train_x = train_x.drop(columns=['id'])\n",
+    "valid_x = valid_x.drop(columns=['id'])\n",
+    "test_x = test_x.drop(columns=['id'])\n",
+    "\n",
+    "print('train_x shape:', train_x.shape)\n",
+    "print('test_x shape:', test_x.shape)\n",
+    "print('valid_x shape:', valid_x.shape)\n",
+    "# print column names\n",
+    "print('features:', train_x.columns.to_list())\n",
+    "feature_names = train_x.columns.to_list()\n",
+    "\n",
+    "# Create an imputer object with a mean filling strategy\n",
+    "imputer = SimpleImputer(strategy='mean')\n",
+    "\n",
+    "train_x = imputer.fit_transform(train_x)\n",
+    "valid_x = imputer.transform(valid_x)\n",
+    "test_x = imputer.transform(test_x)\n",
+    "\n",
+    "# Scale Data between 0 and 1\n",
+    "scaler = MinMaxScaler()\n",
+    "# Fit the scaler to your data and then transform it\n",
+    "train_x = scaler.fit_transform(train_x)\n",
+    "valid_x = scaler.transform(valid_x)\n",
+    "test_x = scaler.transform(test_x)\n",
+    "\n",
+    "\n",
+    "\n",
+    "# use xgboost\n",
+    "dtrain = xgb.DMatrix(train_x, label=train_y)\n",
+    "dvalid = xgb.DMatrix(valid_x, label=valid_y)\n",
+    "dtest = xgb.DMatrix(test_x, label=test_y)\n",
+    "\n",
+    "num_classes= len(set(valid_y.to_list()))\n",
+    "print('number of classes:', num_classes)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Validierungsgenauigkeit: 0.7557077625570776\n",
+      "Testgenauigkeit: 0.7922374429223744\n"
+     ]
+    }
+   ],
+   "source": [
+    "\n",
+    "from sklearn.tree import DecisionTreeClassifier\n",
+    "from sklearn.metrics import accuracy_score\n",
+    "\n",
+    "\n",
+    "# Beispiel: Begrenzung der Tiefe des Baumes\n",
+    "dt_classifier = DecisionTreeClassifier(max_depth=5)\n",
+    "\n",
+    "# Schritt 3: Trainieren des Modells mit Trainingsdaten\n",
+    "dt_classifier.fit(train_x, train_y)\n",
+    "\n",
+    "# Schritt 4: Bewertung des Modells mit Validierungsdaten\n",
+    "valid_pred = dt_classifier.predict(valid_x)\n",
+    "valid_accuracy = accuracy_score(valid_y, valid_pred)\n",
+    "print(f'Validierungsgenauigkeit: {valid_accuracy}')\n",
+    "\n",
+    "# Schritt 5: Hyperparameter-Optimierung\n",
+    "\n",
+    "# Schritt 6: Endgültige Bewertung mit Testdaten\n",
+    "test_pred = dt_classifier.predict(test_x)\n",
+    "test_accuracy = accuracy_score(test_y, test_pred)\n",
+    "print(f'Testgenauigkeit: {test_accuracy}')\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Die Validierungsgenauigkeit des Modells liegt bei 75,5%, was darauf hinweist, dass das Modell in etwa drei Vierteln der Fälle korrekte Vorhersagen auf den Validierungsdaten macht. Dies zeigt eine recht solide Leistung, deutet jedoch auch darauf hin, dass es noch Verbesserungspotenzial gibt, insbesondere bei der Verfeinerung des Modells, um die Fehlerquote zu senken"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "Mit einer Testgenauigkeit von 79% klassifiziert das Modell die Testdaten überwiegend korrekt. Dieses Ergebnis ist ein Indikator dafür, dass das Modell eine gute Generalisierungsfähigkeit aufweist und zuverlässig auf neuen, unbekannten Daten agieren kann. "
+   ]
+  }
+ ],
+ "metadata": {
+  "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": "3.11.9"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}