analyze_results.py

import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.metrics import confusion_matrix, cohen_kappa_score
import numpy as np

def plot_confusion_matrix(conf_matrix, labels, title, filename):
    plt.figure(figsize=(12, 10))
    sns.set(font_scale=0.7)  # 调整字体比例
    sns.heatmap(conf_matrix, annot=True, fmt=".2f", cmap="Blues", 
                cbar_kws={'label': '%'})
    plt.xlabel('predict')
    plt.ylabel('true')
    plt.title(title)
    plt.xticks(ticks=np.arange(len(labels)) + 0.5, labels=labels, rotation=45, ha='right')
    plt.yticks(ticks=np.arange(len(labels)) + 0.5, labels=labels, rotation=0)
    plt.tight_layout()
    
    # 保存图片到指定路径
    plt.savefig(filename, format='png')
    
    # 显示图片
    plt.show()

def process_csv(file_path, title, filename):
    df = pd.read_csv(file_path)
    y_true = df['expected'].astype(str)
    y_pred = df['predicted'].astype(str)

    labels = sorted(set(y_true) | set(y_pred))
    conf_matrix = confusion_matrix(y_true, y_pred, labels=labels, normalize='true') * 100

    kappa = cohen_kappa_score(y_true, y_pred)
    
    print(f"{title} Kappa 系数: {kappa:.2f}")
    plot_confusion_matrix(conf_matrix, labels, title, filename)

first_level_results_file = 'first_level_results_true.csv'
second_level_results_file = 'second_level_results_true.csv'

first_level_image_file = 'first_level_confusion_matrix.png'
second_level_image_file = 'second_level_confusion_matrix.png'

process_csv(first_level_results_file, 'FirstLevel', first_level_image_file)
process_csv(second_level_results_file, 'SecondLevel', second_level_image_file)