Uncovering the Best Email Spam Classifier: A Comparative Study of Machine Learning Algorithms

import pandas as pdimport numpy as npfrom sklearn.model_selection import train_test_splitfrom sklearn.linear_model import LogisticRegressionfrom sklearn.neighbors import KNeighborsClassifierfrom sklearn.svm import SVCfrom sklearn.naive_bayes import GaussianNBfrom sklearn.tree import DecisionTreeClassifierfrom sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifierfrom sklearn.metrics import accuracy_score, precision_score, recall_score, f1_scoreimport matplotlib.pyplot as plt

# Load the data from csv file into a pandas dataframedf = pd.read_csv("/kaggle/input/email-spam-classification-dataset-csv/emails.csv")# Remove the first column (Email name) as it is not relevant for the predictiondf = df.drop(columns=['Email No.'])# Replace non-numeric characters with NaN valuesdf = df.replace(r'[^\d.]+', value=float('nan'), regex=True)# Fill missing values with 0df.fillna(0, inplace=True)# Convert the data into integer typedf = df.astype(int)df

# Split the data into features (X) and labels (y)X = df.iloc[:, :-1].valuesy = df.iloc[:, -1].values

# Split the data into training and testing setsX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)

# Define a dictionary to store the results of each modelresults = {}

# Train Logistic Regression modelmodel_LR = LogisticRegression()model_LR.fit(X_train, y_train)# Predict the target values for test sety_pred_LR = model_LR.predict(X_test)# Evaluate the Logistic Regression modelaccuracy_LR = accuracy_score(y_test, y_pred_LR)precision_LR = precision_score(y_test, y_pred_LR)recall_LR = recall_score(y_test, y_pred_LR)f1_LR = f1_score(y_test, y_pred_LR)# Store the results of Logistic Regression model in the dictionaryresults["Logistic Regression"] = {"accuracy": accuracy_LR,                                   "precision": precision_LR,                                   "recall": recall_LR,                                   "f1_score": f1_LR                                 }

# Train KNN modelmodel_KNN = KNeighborsClassifier()model_KNN.fit(X_train, y_train)# Predict the target values for test sety_pred_KNN = model_KNN.predict(X_test)# Evaluate the KNN modelaccuracy_KNN = accuracy_score(y_test, y_pred_KNN)precision_KNN = precision_score(y_test, y_pred_KNN)recall_KNN = recall_score(y_test, y_pred_KNN)f1_KNN = f1_score(y_test, y_pred_KNN)# Store the results of KNN model in the dictionaryresults["KNN"] = {"accuracy": accuracy_KNN,                   "precision": precision_KNN,                   "recall": recall_KNN,                   "f1_score": f1_KNN                 }

# Train SVM modelmodel_SVM = SVC()model_SVM.fit(X_train, y_train)# Predict the target values for test sety_pred_SVM = model_SVM.predict(X_test)# Evaluate the SVM modelaccuracy_SVM = accuracy_score(y_test, y_pred_SVM)precision_SVM = precision_score(y_test, y_pred_SVM)recall_SVM = recall_score(y_test, y_pred_SVM)f1_SVM = f1_score(y_test, y_pred_SVM)# Store the results of SVM model in the dictionaryresults["SVM"] = {"accuracy": accuracy_SVM,                   "precision": precision_SVM,                   "recall": recall_SVM,                   "f1_score": f1_SVM                 }

# Train Naive Bayes modelmodel_NB = GaussianNB()model_NB.fit(X_train, y_train)# Predict the target values for test sety_pred_NB = model_NB.predict(X_test)# Evaluate the Naive Bayes modelaccuracy_NB = accuracy_score(y_test, y_pred_NB)precision_NB = precision_score(y_test, y_pred_NB)recall_NB = recall_score(y_test, y_pred_NB)f1_NB = f1_score(y_test, y_pred_NB)# Store the results of Naive Bayes model in the dictionaryresults["Naive Bayes"] = {"accuracy": accuracy_NB,                           "precision": precision_NB,                           "recall": recall_NB,                           "f1_score": f1_NB                         }

# Train Decision Tree modelmodel_DT = DecisionTreeClassifier()model_DT.fit(X_train, y_train)# Predict the target values for test sety_pred_DT = model_DT.predict(X_test)# Evaluate the Decision Tree modelaccuracy_DT = accuracy_score(y_test, y_pred_DT)precision_DT = precision_score(y_test, y_pred_DT)recall_DT = recall_score(y_test, y_pred_DT)f1_DT = f1_score(y_test, y_pred_DT)# Store the results of Decision Tree model in the dictionaryresults["Decision Tree"] = {"accuracy": accuracy_DT,                             "precision": precision_DT,                             "recall": recall_DT,                             "f1_score": f1_DT                           }

# Train Random Forest modelmodel_RF = RandomForestClassifier()model_RF.fit(X_train, y_train)# Predict the target values for test sety_pred_RF = model_RF.predict(X_test)# Evaluate the Random Forest modelaccuracy_RF = accuracy_score(y_test, y_pred_RF)precision_RF = precision_score(y_test, y_pred_RF)recall_RF = recall_score(y_test, y_pred_RF)f1_RF = f1_score(y_test, y_pred_RF)# Store the results of Random Forest model in the dictionaryresults["Random Forest"] = {"accuracy": accuracy_RF,                             "precision": precision_RF,                             "recall": recall_RF,                             "f1_score": f1_RF                           }

# Train Logistic Regression modelmodel_LR = LogisticRegression()model_LR.fit(X_train, y_train)# Predict the target values for test sety_pred_LR = model_LR.predict(X_test)# Evaluate the Logistic Regression modelaccuracy_LR = accuracy_score(y_test, y_pred_LR)precision_LR = precision_score(y_test, y_pred_LR)recall_LR = recall_score(y_test, y_pred_LR)f1_LR = f1_score(y_test, y_pred_LR)# Store the results of Logistic Regression model in the dictionaryresults["Logistic Regression"] = {"accuracy": accuracy_LR,                                   "precision": precision_LR,                                   "recall": recall_LR,                                   "f1_score": f1_LR                                 }

# Print the final results of all modelsprint(results)

{'Logistic Regression': {'accuracy': 0.9652173913043478,  'precision': 0.9220338983050848,  'recall': 0.9543859649122807,  'f1_score': 0.9379310344827585}, 'KNN': {'accuracy': 0.8541062801932368,  'precision': 0.7018072289156626,  'recall': 0.8175438596491228,  'f1_score': 0.7552674230145866}, 'SVM': {'accuracy': 0.7951690821256039,  'precision': 0.8067226890756303,  'recall': 0.3368421052631579,  'f1_score': 0.4752475247524753}, 'Naive Bayes': {'accuracy': 0.9584541062801932,  'precision': 0.8853503184713376,  'recall': 0.9754385964912281,  'f1_score': 0.9282136894824708}, 'Decision Tree': {'accuracy': 0.9285024154589372,  'precision': 0.8625429553264605,  'recall': 0.8807017543859649,  'f1_score': 0.8715277777777777}, 'Random Forest': {'accuracy': 0.9652173913043478,  'precision': 0.9368421052631579,  'recall': 0.9368421052631579,  'f1_score': 0.9368421052631579}}

# Plot the accuracy of all modelsplt.figure(figsize=(10,5))plt.bar(results.keys(), [result["accuracy"] for result in results.values()])plt.title("Accuracy of Different Models")plt.xlabel("Models")plt.ylabel("Accuracy")plt.ylim(0, 1)plt.show()

# Plot the precision of all modelsplt.figure(figsize=(10,5))plt.bar(results.keys(), [result["precision"] for result in results.values()])plt.title("Precision of Different Models")plt.xlabel("Models")plt.ylabel("Precision")plt.ylim(0, 1)plt.show()

# Plot the recall of all modelsplt.figure(figsize=(10,5))plt.bar(results.keys(), [result["recall"] for result in results.values()])plt.title("Recall of Different Models")plt.xlabel("Models")plt.ylabel("Recall")plt.ylim(0, 1)plt.show()

# Plot the F1-score of all modelsplt.figure(figsize=(10,5))plt.bar(results.keys(), [result["f1_score"] for result in results.values()])plt.title("F1-Score of Different Models")plt.xlabel("Models")plt.ylabel("F1-Score")plt.ylim(0, 1)plt.show()

# extract metrics valuesaccuracies = [results[model]['accuracy'] for model in results]precisions = [results[model]['precision'] for model in results]recalls = [results[model]['recall'] for model in results]f1_scores = [results[model]['f1_score'] for model in results]# plot bar chartbar_width = 0.2index = np.arange(len(results))plt.bar(index, accuracies, bar_width, label='Accuracy')plt.bar(index + bar_width, precisions, bar_width, label='Precision')plt.bar(index + 2 * bar_width, recalls, bar_width, label='Recall')plt.bar(index + 3 * bar_width, f1_scores, bar_width, label='F1 Score')plt.xticks(index + bar_width, list(results.keys()))plt.legend()plt.show()