SpamDetective: Spam Classification Using Naive Bayes Models

import warnings
warnings.filterwarnings('ignore')
import streamlit as st
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sn
import pickle
from sklearn.metrics import confusion_matrix, classification_report
from wordcloud import WordCloud
from collections import Counter
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.model_selection import train_test_split
from sklearn.naive_bayes import GaussianNB, BernoulliNB, ComplementNB, MultinomialNB
from imblearn.over_sampling import SMOTE
df=pd.read_csv("C:/Users/diva1/Downloads/spam (1).csv")
df['spam']=df['Category'].apply(lambda x:1 if x=='spam' else 0)
st.sidebar.title("Choose a Model")
model_choice=st.sidebar.selectbox(
    "Select a model:",
    ["MultinomialNB","GaussianNB","BernoulliNB","ComplementNB"]
)
vect=TfidfVectorizer()
X=vect.fit_transform(df['Message'])
Y=df['spam']
smote = SMOTE()
X_resampled,y_resampled=smote.fit_resample(X,Y)
X_train,X_test,y_train,y_test=train_test_split(X_resampled,y_resampled,test_size=0.2)
if model_choice=="MultinomialNB":
    model=MultinomialNB(alpha=1)
elif model_choice=="GaussianNB":
    model=GaussianNB()
elif model_choice=="BernoulliNB":
    model=BernoulliNB(alpha=1)
elif model_choice=="ComplementNB":
    model=ComplementNB(alpha=1)
model.fit(X_train.toarray(),y_train)
y_pred=model.predict(X_test.toarray())
cm=confusion_matrix(y_test,y_pred)
clf_rep=classification_report(y_test,y_pred,output_dict=True)
df_report=pd.DataFrame(clf_rep).transpose()
st.title(f'{model_choice}-Spam Detection')
st.subheader('Confusion Matrix')
fig, ax = plt.subplots(figsize=(8,6)) 
sn.heatmap(cm, annot=True, fmt='d', cmap='Blues', xticklabels=['Ham', 'Spam'], yticklabels=['Ham', 'Spam'], ax=ax)
ax.set_title('Confusion Matrix')
st.pyplot(fig)
st.subheader('Classification Report')
st.write(df_report)
st.subheader('Word Clouds')
for category in df['Category'].unique():
    text=' '.join(df[df['Category']==category]['Message'])
    wordcloud=WordCloud(width=600,height=400,background_color='white').generate(text)
    st.image(wordcloud.to_array(),caption=f"Word Cloud for {category}")
if st.button('Save Model'):
    with open(f'spam_model_{model_choice}.pickle','wb') as f:
        pickle.dump(model,f)
    st.success(f"Model {model_choice} saved successfully!")
st.subheader('Test on Sample Emails')
emails=st.text_area("Enter email content here (separate emails by newline):").split('\n')
if emails:
    emails_count=vect.transform(emails)
    predictions=model.predict(emails_count.toarray())
    st.write("Predictions (0 = Ham, 1 = Spam):",predictions)