app.py

import streamlit as st
import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import StandardScaler
from sklearn.ensemble import VotingClassifier, StackingClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.tree import DecisionTreeClassifier
from sklearn.ensemble import RandomForestClassifier, GradientBoostingClassifier
from sklearn.svm import SVC
from sklearn.metrics import accuracy_score, confusion_matrix, roc_curve, auc, classification_report
import matplotlib.pyplot as plt
import seaborn as sns

# 設定 Streamlit 介面標題
st.title('分類模型比較：堆疊與投票分類器')

# 讓使用者上傳資料
uploaded_file = st.file_uploader("請上傳 CSV 檔案", type=["csv"])

if uploaded_file is not None:
    df = pd.read_csv(uploaded_file)

    # 定義特徵與目標變數
    X = df.drop(columns=['Target_goal'])
    y = df['Target_goal']

    # 分割數據集
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

    # 標準化數據
    scaler = StandardScaler()
    X_train = scaler.fit_transform(X_train)
    X_test = scaler.transform(X_test)

    # 定義基礎模型
    estimators = [
        ('lr', LogisticRegression()),
        ('dt', DecisionTreeClassifier()),
        ('rf', RandomForestClassifier()),
        ('gb', GradientBoostingClassifier()),
        ('svc', SVC(probability=True))
    ]

    # 堆疊分類器
    stacking_clf = StackingClassifier(
        estimators=estimators,
        final_estimator=LogisticRegression()
    )
    stacking_clf.fit(X_train, y_train)
    y_pred_stack = stacking_clf.predict(X_test)
    y_pred_stack_proba = stacking_clf.predict_proba(X_test)[:, 1]
    
    # 堆疊分類器準確性
    accuracy_stack = accuracy_score(y_test, y_pred_stack)
    st.write(f'堆疊分類器的準確性: {accuracy_stack:.2f}')
    
    # 堆疊分類器的分類報告
    st.write("堆疊分類器的分類報告:")
    st.text(classification_report(y_test, y_pred_stack))

    # 投票分類器
    voting_clf = VotingClassifier(
        estimators=estimators,
        voting='soft'
    )
    voting_clf.fit(X_train, y_train)
    y_pred_vote = voting_clf.predict(X_test)
    y_pred_vote_proba = voting_clf.predict_proba(X_test)[:, 1]
    
    # 投票分類器準確性
    accuracy_vote = accuracy_score(y_test, y_pred_vote)
    st.write(f'投票分類器的準確性: {accuracy_vote:.2f}')
    
    # 投票分類器的分類報告
    st.write("投票分類器的分類報告:")
    st.text(classification_report(y_test, y_pred_vote))

    # 混淆矩陣可視化
    st.write("堆疊分類器的混淆矩陣：")
    conf_matrix_stack = confusion_matrix(y_test, y_pred_stack)
    fig, ax = plt.subplots()
    sns.heatmap(conf_matrix_stack, annot=True, fmt='d', cmap='Blues', ax=ax)
    ax.set_title('堆疊分類器的混淆矩陣')
    st.pyplot(fig)

    st.write("投票分類器的混淆矩陣：")
    conf_matrix_vote = confusion_matrix(y_test, y_pred_vote)
    fig, ax = plt.subplots()
    sns.heatmap(conf_matrix_vote, annot=True, fmt='d', cmap='Blues', ax=ax)
    ax.set_title('投票分類器的混淆矩陣')
    st.pyplot(fig)

    # ROC 曲線
    fpr_stack, tpr_stack, _ = roc_curve(y_test, y_pred_stack_proba)
    roc_auc_stack = auc(fpr_stack, tpr_stack)

    fpr_vote, tpr_vote, _ = roc_curve(y_test, y_pred_vote_proba)
    roc_auc_vote = auc(fpr_vote, tpr_vote)

    fig, ax = plt.subplots()
    ax.plot(fpr_stack, tpr_stack, color='blue', lw=2, label='堆疊分類器 (AUC = %0.2f)' % roc_auc_stack)
    ax.plot(fpr_vote, tpr_vote, color='red', lw=2, label='投票分類器 (AUC = %0.2f)' % roc_auc_vote)
    ax.plot([0, 1], [0, 1], color='gray', lw=1, linestyle='--')
    ax.set_xlim([0.0, 1.0])
    ax.set_ylim([0.0, 1.05])
    ax.set_xlabel('假陽性率（False Positive Rate）')
    ax.set_ylabel('真陽性率（True Positive Rate）')
    ax.set_title('ROC 曲線')
    ax.legend(loc="lower right")
    st.pyplot(fig)