Classification-Algorithm.md

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt   ##绘图库

plt.rcParams['font.sans-serif'] = ['SimHei']  # 指定默认字体为新宋体。
plt.rcParams['axes.unicode_minus'] = False  # 解决保存图像时'-'显示为方块的问题。

from sklearn.model_selection import train_test_split   ## 划分

import torch

from sklearn.model_selection import train_test_split   #训练集测试集划分
from sklearn.ensemble import RandomForestClassifier    #随机森林相关库
from sklearn.metrics import accuracy_score, recall_score, precision_score, f1_score        #模型精度评分
from sklearn.metrics import confusion_matrix           #混淆矩阵表
from sklearn import svm  #支持向量机

# 检查是否有可用的 GPU
if torch.cuda.is_available():
    device = torch.device("cuda")  # 使用GPU
    print("Using GPU:", torch.cuda.get_device_name(0))
else:
    device = torch.device("cpu")  # 使用 CPU
    print("Using CPU")

Using GPU: NVIDIA GeForce MX450

# 加载 csv文件
merged_data = pd.read_csv('merged_data.csv')

data = merged_data.drop(['AQI','PM2.5','PM10','NO2','CO','SO2','O3_8h'],axis=1)
# 检查结果
data

<style scoped> .dataframe tbody tr th:only-of-type { vertical-align: middle; }

.dataframe tbody tr th {
    vertical-align: top;
}

.dataframe thead th {
    text-align: right;
}

</style>

	日期	气温	气压	平均海平面大气压	气压趋势	湿度	风向	平均风速	最大阵风	天气情况	最低气温	最高气温	露点温度	降水时间	季节	质量等级
0	2013-12-02	3.162	759.712	764.462	-0.013	45.875	7.375	1.375	7.177	5.500	6.896	18.301	-8.588	10.717	冬	轻度污染
1	2013-12-03	5.488	761.725	766.425	0.100	39.000	8.000	1.625	7.177	4.750	7.209	18.314	-8.900	10.717	冬	良
2	2013-12-04	5.250	760.300	764.988	-0.138	45.375	9.375	1.250	7.177	1.750	7.134	18.714	-6.675	10.717	冬	轻度污染
3	2013-12-05	6.150	763.275	767.975	0.250	30.000	6.875	2.250	7.177	3.875	7.759	18.551	-10.912	10.717	冬	良
4	2013-12-06	2.925	760.325	765.075	-0.275	52.750	4.875	1.250	7.177	1.000	7.121	18.239	-6.350	10.717	冬	中度污染
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
3676	2023-12-26	-2.050	770.638	773.938	0.925	54.125	8.250	1.500	7.147	0.500	-9.475	2.588	-10.862	12.000	冬	良
3677	2023-12-27	-3.888	771.538	774.850	-0.538	67.750	7.250	1.125	4.250	0.000	-8.900	5.348	-9.450	12.000	冬	良
3678	2023-12-28	-3.012	769.138	772.438	-0.038	69.875	6.625	1.000	3.750	0.375	-9.100	3.750	-8.288	12.000	冬	轻度污染
3679	2023-12-29	-2.800	765.112	768.400	-0.938	78.125	5.625	1.125	4.147	2.000	-6.302	3.975	-6.300	12.000	冬	轻度污染
3680	2023-12-30	-1.238	760.250	763.512	0.225	75.125	2.625	1.125	3.875	2.000	-6.562	2.950	-5.625	12.000	冬	优

3681 rows × 16 columns

import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.preprocessing import LabelEncoder

# 初始化标签编码器
le = LabelEncoder()

# 对质量等级进行编码
data['质量等级'] = le.fit_transform(data['质量等级'])
data['季节'] = le.fit_transform(data['季节'])

# 选择特定的特征列
selected_features = ['气温', '气压', '平均海平面大气压', '气压趋势', '湿度', '风向', '平均风速', '最大阵风', '天气情况', '最低气温', '最高气温', '露点温度', '降水时间', '季节', '质量等级']

# 计算选定特征列的相关性
correlation = data[selected_features].corr()['质量等级']

correlation = correlation.abs()

# 按照绝对值降序排序
correlation = correlation.reindex(correlation.abs().sort_values(ascending=False).index)

# 创建一个条形图
plt.figure(figsize=(12, 8))
sns.barplot(x=correlation[1:], hue = correlation.index[1:], legend = 'auto', palette='coolwarm')

# 添加标题和标签
plt.title('质量等级的相关性分析')
plt.xlabel('相关性')
plt.ylabel('特征值')

# 显示图表
plt.show()

print(correlation)

质量等级        1.000000
平均风速        0.164287
最大阵风        0.113761
最高气温        0.111750
气压          0.109454
平均海平面大气压    0.106138
露点温度        0.098350
气温          0.095311
气压趋势        0.079601
最低气温        0.076747
风向          0.070087
天气情况        0.067813
降水时间        0.063081
湿度          0.052930
季节          0.043512
Name: 质量等级, dtype: float64

划分训练集，尽可能地减少对难预测的空气污染物成分的依赖

# 将data的值复制到df当中
df = merged_data


# 执行独热编码转换类别字段
df = pd.get_dummies(df, columns=['季节'])


# 预测前，将数据集划分为训练集和验证集，尽可能地减少对难预测的空气污染物成分的依赖
X = df.drop(columns=['质量等级','日期','AQI','O3_8h','NO2','SO2'])
y = df['质量等级']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

使用随机森林的默认状态进行调试

随机森林参数调优

# 使用随机森林分类器进行训练
# classifier = RandomForestClassifier(n_estimators=120, random_state=42)
## 参数调优 
classifier = RandomForestClassifier(n_estimators=50, 
                                    min_samples_leaf=5, 
                                    min_samples_split=10, 
                                    random_state=42)

classifier.fit(X_train, y_train)

# 进行预测并检查准确率
predictions = classifier.predict(X_test)
accuracy = accuracy_score(y_test, predictions)
print("预测的准确率是：", accuracy)
# 概率
predicted_proba = classifier.predict_proba(X_test)

# 计算混淆矩阵并创建热力图
cm = confusion_matrix(y_test, predictions)
plt.figure(figsize=(10, 7))
sns.heatmap(cm, annot = True, cmap = 'Blues')
plt.title('随机森林方法')
plt.xlabel('预测值')
plt.ylabel('真实值')



# 训练集上的预测
train_predictions = classifier.predict(X_train)
# 计算准确率、召回率、精确率 和 F1分数
accuracy_train = classifier.score(X_train, y_train)
recall_train = recall_score(y_train, train_predictions, average='weighted')
precision_train = precision_score(y_train, train_predictions, average='weighted',zero_division=1)
f1_train = f1_score(y_train, train_predictions, average='weighted')
accuracy_test = accuracy
recall_test = recall_score(y_test, predictions, average='weighted')
precision_test = precision_score(y_test, predictions, average='weighted',zero_division=1)
f1_test = f1_score(y_test, predictions, average='weighted')
# 创建 DataFrame
performance = pd.DataFrame({
    '准确率': [accuracy_train, accuracy_test],
    '召回率': [recall_train, recall_test],
    '精确率': [precision_train, precision_test],
    'F1': [f1_train, f1_test]
}, index = ['训练集', '测试集'])
# 显示 performance
performance_styler = performance.style.set_properties(**{'text-align': 'center'})
display(performance_styler)
# 构建预测结果对照表
results = pd.DataFrame({
    '真实值': y_test,
    '预测值': predictions
})

# 获得类别列表，按照模型内部的顺序
class_list = classifier.classes_

# 将预测的概率与其对应的类别关联起来
for i, quality_level in enumerate(class_list):
    results[f'{quality_level}预测概率'] = predicted_proba[:, i]

# 使用 .head() 方法获取前20条数据
results_head = results.head(20)

# 设置数据显示为居中格式
results_styler = results_head.style.set_properties(**{'text-align': 'center'})

# 显示居中对齐的前100条数据
display(results_styler)

预测的准确率是： 0.8168249660786974

<style type="text/css"> #T_5c29b_row0_col0, #T_5c29b_row0_col1, #T_5c29b_row0_col2, #T_5c29b_row0_col3, #T_5c29b_row1_col0, #T_5c29b_row1_col1, #T_5c29b_row1_col2, #T_5c29b_row1_col3 { text-align: center; } </style>

	准确率	召回率	精确率	F1
训练集	0.920177	0.920177	0.921693	0.918641
测试集	0.816825	0.816825	0.819219	0.815129

<style type="text/css"> #T_93ae6_row0_col0, #T_93ae6_row0_col1, #T_93ae6_row0_col2, #T_93ae6_row0_col3, #T_93ae6_row0_col4, #T_93ae6_row0_col5, #T_93ae6_row0_col6, #T_93ae6_row0_col7, #T_93ae6_row0_col8, #T_93ae6_row1_col0, #T_93ae6_row1_col1, #T_93ae6_row1_col2, #T_93ae6_row1_col3, #T_93ae6_row1_col4, #T_93ae6_row1_col5, #T_93ae6_row1_col6, #T_93ae6_row1_col7, #T_93ae6_row1_col8, #T_93ae6_row2_col0, #T_93ae6_row2_col1, #T_93ae6_row2_col2, #T_93ae6_row2_col3, #T_93ae6_row2_col4, #T_93ae6_row2_col5, #T_93ae6_row2_col6, #T_93ae6_row2_col7, #T_93ae6_row2_col8, #T_93ae6_row3_col0, #T_93ae6_row3_col1, #T_93ae6_row3_col2, #T_93ae6_row3_col3, #T_93ae6_row3_col4, #T_93ae6_row3_col5, #T_93ae6_row3_col6, #T_93ae6_row3_col7, #T_93ae6_row3_col8, #T_93ae6_row4_col0, #T_93ae6_row4_col1, #T_93ae6_row4_col2, #T_93ae6_row4_col3, #T_93ae6_row4_col4, #T_93ae6_row4_col5, #T_93ae6_row4_col6, #T_93ae6_row4_col7, #T_93ae6_row4_col8, #T_93ae6_row5_col0, #T_93ae6_row5_col1, #T_93ae6_row5_col2, #T_93ae6_row5_col3, #T_93ae6_row5_col4, #T_93ae6_row5_col5, #T_93ae6_row5_col6, #T_93ae6_row5_col7, #T_93ae6_row5_col8, #T_93ae6_row6_col0, #T_93ae6_row6_col1, #T_93ae6_row6_col2, #T_93ae6_row6_col3, #T_93ae6_row6_col4, #T_93ae6_row6_col5, #T_93ae6_row6_col6, #T_93ae6_row6_col7, #T_93ae6_row6_col8, #T_93ae6_row7_col0, #T_93ae6_row7_col1, #T_93ae6_row7_col2, #T_93ae6_row7_col3, #T_93ae6_row7_col4, #T_93ae6_row7_col5, #T_93ae6_row7_col6, #T_93ae6_row7_col7, #T_93ae6_row7_col8, #T_93ae6_row8_col0, #T_93ae6_row8_col1, #T_93ae6_row8_col2, #T_93ae6_row8_col3, #T_93ae6_row8_col4, #T_93ae6_row8_col5, #T_93ae6_row8_col6, #T_93ae6_row8_col7, #T_93ae6_row8_col8, #T_93ae6_row9_col0, #T_93ae6_row9_col1, #T_93ae6_row9_col2, #T_93ae6_row9_col3, #T_93ae6_row9_col4, #T_93ae6_row9_col5, #T_93ae6_row9_col6, #T_93ae6_row9_col7, #T_93ae6_row9_col8, #T_93ae6_row10_col0, #T_93ae6_row10_col1, #T_93ae6_row10_col2, #T_93ae6_row10_col3, #T_93ae6_row10_col4, #T_93ae6_row10_col5, #T_93ae6_row10_col6, #T_93ae6_row10_col7, #T_93ae6_row10_col8, #T_93ae6_row11_col0, #T_93ae6_row11_col1, #T_93ae6_row11_col2, #T_93ae6_row11_col3, #T_93ae6_row11_col4, #T_93ae6_row11_col5, #T_93ae6_row11_col6, #T_93ae6_row11_col7, #T_93ae6_row11_col8, #T_93ae6_row12_col0, #T_93ae6_row12_col1, #T_93ae6_row12_col2, #T_93ae6_row12_col3, #T_93ae6_row12_col4, #T_93ae6_row12_col5, #T_93ae6_row12_col6, #T_93ae6_row12_col7, #T_93ae6_row12_col8, #T_93ae6_row13_col0, #T_93ae6_row13_col1, #T_93ae6_row13_col2, #T_93ae6_row13_col3, #T_93ae6_row13_col4, #T_93ae6_row13_col5, #T_93ae6_row13_col6, #T_93ae6_row13_col7, #T_93ae6_row13_col8, #T_93ae6_row14_col0, #T_93ae6_row14_col1, #T_93ae6_row14_col2, #T_93ae6_row14_col3, #T_93ae6_row14_col4, #T_93ae6_row14_col5, #T_93ae6_row14_col6, #T_93ae6_row14_col7, #T_93ae6_row14_col8, #T_93ae6_row15_col0, #T_93ae6_row15_col1, #T_93ae6_row15_col2, #T_93ae6_row15_col3, #T_93ae6_row15_col4, #T_93ae6_row15_col5, #T_93ae6_row15_col6, #T_93ae6_row15_col7, #T_93ae6_row15_col8, #T_93ae6_row16_col0, #T_93ae6_row16_col1, #T_93ae6_row16_col2, #T_93ae6_row16_col3, #T_93ae6_row16_col4, #T_93ae6_row16_col5, #T_93ae6_row16_col6, #T_93ae6_row16_col7, #T_93ae6_row16_col8, #T_93ae6_row17_col0, #T_93ae6_row17_col1, #T_93ae6_row17_col2, #T_93ae6_row17_col3, #T_93ae6_row17_col4, #T_93ae6_row17_col5, #T_93ae6_row17_col6, #T_93ae6_row17_col7, #T_93ae6_row17_col8, #T_93ae6_row18_col0, #T_93ae6_row18_col1, #T_93ae6_row18_col2, #T_93ae6_row18_col3, #T_93ae6_row18_col4, #T_93ae6_row18_col5, #T_93ae6_row18_col6, #T_93ae6_row18_col7, #T_93ae6_row18_col8, #T_93ae6_row19_col0, #T_93ae6_row19_col1, #T_93ae6_row19_col2, #T_93ae6_row19_col3, #T_93ae6_row19_col4, #T_93ae6_row19_col5, #T_93ae6_row19_col6, #T_93ae6_row19_col7, #T_93ae6_row19_col8 { text-align: center; } </style>

	真实值	预测值	严重污染预测概率	中度污染预测概率	优预测概率	无预测概率	良预测概率	轻度污染预测概率	重度污染预测概率
1097	严重污染	重度污染	0.242753	0.028671	0.000000	0.000000	0.007273	0.003818	0.717485
2784	优	优	0.002857	0.003500	0.580033	0.000000	0.355595	0.058015	0.000000
2440	良	良	0.000000	0.035993	0.159627	0.002500	0.588910	0.212970	0.000000
1694	优	优	0.001538	0.008901	0.475104	0.000000	0.389171	0.119131	0.006154
2494	良	良	0.000000	0.000000	0.163351	0.002500	0.757525	0.076624	0.000000
2270	轻度污染	轻度污染	0.000000	0.168943	0.012083	0.000000	0.035368	0.750467	0.033139
3477	良	良	0.000000	0.008205	0.188131	0.001250	0.676005	0.125076	0.001333
937	良	中度污染	0.000000	0.530691	0.000000	0.000000	0.141202	0.304054	0.024053
495	中度污染	中度污染	0.064650	0.449891	0.012000	0.000000	0.082585	0.196608	0.194265
798	重度污染	重度污染	0.120026	0.219474	0.015000	0.000000	0.092429	0.229477	0.323595
3375	优	优	0.001667	0.000000	0.980167	0.000000	0.016500	0.001667	0.000000
1747	优	优	0.006000	0.000000	0.726175	0.005778	0.256968	0.005079	0.000000
1487	重度污染	重度污染	0.151462	0.156333	0.000000	0.000000	0.003333	0.012222	0.676650
969	中度污染	中度污染	0.005934	0.564135	0.003205	0.000000	0.034687	0.221312	0.170726
2883	中度污染	中度污染	0.016024	0.445166	0.000000	0.000000	0.099437	0.374235	0.065138
655	轻度污染	轻度污染	0.000000	0.108666	0.000000	0.000000	0.220001	0.654238	0.017095
229	轻度污染	轻度污染	0.001538	0.299135	0.002857	0.000000	0.083248	0.609888	0.003333
1116	良	良	0.000000	0.013553	0.084372	0.000000	0.880953	0.016122	0.005000
840	中度污染	中度污染	0.020206	0.757988	0.000000	0.000000	0.004524	0.162131	0.055151
32	良	良	0.000000	0.016905	0.008667	0.000000	0.895619	0.078810	0.000000

支持向量机分类算法

# 使用支持向量机分类器进行训练
# 默认参数
# classifier = svm.SVC(probability=True)
# classifier.fit(X_train, y_train)


# 参数调优
classifier = svm.SVC(probability=True, C=0.8, kernel='linear', gamma=0.01)
classifier.fit(X_train, y_train)

# 执行预测并计算准确度
predictions = classifier.predict(X_test)
accuracy = accuracy_score(y_test, predictions)
print("预测准确率：", accuracy)

# 概率
predicted_proba = classifier.predict_proba(X_test)

# 进行预测并检查准确率
predictions = classifier.predict(X_test)
accuracy = accuracy_score(y_test, predictions)
print("预测的准确率是：", accuracy)
# 概率
predicted_proba = classifier.predict_proba(X_test)

# 计算混淆矩阵并创建热力图
cm = confusion_matrix(y_test, predictions)
plt.figure(figsize=(10, 7))
sns.heatmap(cm, annot = True, cmap = 'Blues')
plt.title('支持向量机方法')
plt.xlabel('预测值')
plt.ylabel('真实值')



# 训练集上的预测
train_predictions = classifier.predict(X_train)
# 计算准确率、召回率、精确率 和 F1分数
accuracy_train = classifier.score(X_train, y_train)
recall_train = recall_score(y_train, train_predictions, average='weighted')
precision_train = precision_score(y_train, train_predictions, average='weighted',zero_division=1)
f1_train = f1_score(y_train, train_predictions, average='weighted')
accuracy_test = accuracy
recall_test = recall_score(y_test, predictions, average='weighted')
precision_test = precision_score(y_test, predictions, average='weighted',zero_division=1)
f1_test = f1_score(y_test, predictions, average='weighted')
# 创建 DataFrame
performance = pd.DataFrame({
    '准确率': [accuracy_train, accuracy_test],
    '召回率': [recall_train, recall_test],
    '精确率': [precision_train, precision_test],
    'F1': [f1_train, f1_test]
}, index = ['训练集', '测试集'])
# 显示 performance
performance_styler = performance.style.set_properties(**{'text-align': 'center'})
display(performance_styler)
# 构建预测结果对照表
results = pd.DataFrame({
    '真实值': y_test,
    '预测值': predictions
})

# 获得类别列表，按照模型内部的顺序
class_list = classifier.classes_

# 将预测的概率与其对应的类别关联起来
for i, quality_level in enumerate(class_list):
    results[f'{quality_level}预测概率'] = predicted_proba[:, i]

# 使用 .head() 方法获取前20条数据
results_head = results.head(20)

# 设置数据显示为居中格式
results_styler = results_head.style.set_properties(**{'text-align': 'center'})

# 显示居中对齐的前100条数据
display(results_styler)

预测准确率： 0.7584803256445047
预测的准确率是： 0.7584803256445047

<style type="text/css"> #T_fc809_row0_col0, #T_fc809_row0_col1, #T_fc809_row0_col2, #T_fc809_row0_col3, #T_fc809_row1_col0, #T_fc809_row1_col1, #T_fc809_row1_col2, #T_fc809_row1_col3 { text-align: center; } </style>

	准确率	召回率	精确率	F1
训练集	0.754416	0.754416	0.753321	0.751580
测试集	0.758480	0.758480	0.757210	0.755971

<style type="text/css"> #T_0b4e1_row0_col0, #T_0b4e1_row0_col1, #T_0b4e1_row0_col2, #T_0b4e1_row0_col3, #T_0b4e1_row0_col4, #T_0b4e1_row0_col5, #T_0b4e1_row0_col6, #T_0b4e1_row0_col7, #T_0b4e1_row0_col8, #T_0b4e1_row1_col0, #T_0b4e1_row1_col1, #T_0b4e1_row1_col2, #T_0b4e1_row1_col3, #T_0b4e1_row1_col4, #T_0b4e1_row1_col5, #T_0b4e1_row1_col6, #T_0b4e1_row1_col7, #T_0b4e1_row1_col8, #T_0b4e1_row2_col0, #T_0b4e1_row2_col1, #T_0b4e1_row2_col2, #T_0b4e1_row2_col3, #T_0b4e1_row2_col4, #T_0b4e1_row2_col5, #T_0b4e1_row2_col6, #T_0b4e1_row2_col7, #T_0b4e1_row2_col8, #T_0b4e1_row3_col0, #T_0b4e1_row3_col1, #T_0b4e1_row3_col2, #T_0b4e1_row3_col3, #T_0b4e1_row3_col4, #T_0b4e1_row3_col5, #T_0b4e1_row3_col6, #T_0b4e1_row3_col7, #T_0b4e1_row3_col8, #T_0b4e1_row4_col0, #T_0b4e1_row4_col1, #T_0b4e1_row4_col2, #T_0b4e1_row4_col3, #T_0b4e1_row4_col4, #T_0b4e1_row4_col5, #T_0b4e1_row4_col6, #T_0b4e1_row4_col7, #T_0b4e1_row4_col8, #T_0b4e1_row5_col0, #T_0b4e1_row5_col1, #T_0b4e1_row5_col2, #T_0b4e1_row5_col3, #T_0b4e1_row5_col4, #T_0b4e1_row5_col5, #T_0b4e1_row5_col6, #T_0b4e1_row5_col7, #T_0b4e1_row5_col8, #T_0b4e1_row6_col0, #T_0b4e1_row6_col1, #T_0b4e1_row6_col2, #T_0b4e1_row6_col3, #T_0b4e1_row6_col4, #T_0b4e1_row6_col5, #T_0b4e1_row6_col6, #T_0b4e1_row6_col7, #T_0b4e1_row6_col8, #T_0b4e1_row7_col0, #T_0b4e1_row7_col1, #T_0b4e1_row7_col2, #T_0b4e1_row7_col3, #T_0b4e1_row7_col4, #T_0b4e1_row7_col5, #T_0b4e1_row7_col6, #T_0b4e1_row7_col7, #T_0b4e1_row7_col8, #T_0b4e1_row8_col0, #T_0b4e1_row8_col1, #T_0b4e1_row8_col2, #T_0b4e1_row8_col3, #T_0b4e1_row8_col4, #T_0b4e1_row8_col5, #T_0b4e1_row8_col6, #T_0b4e1_row8_col7, #T_0b4e1_row8_col8, #T_0b4e1_row9_col0, #T_0b4e1_row9_col1, #T_0b4e1_row9_col2, #T_0b4e1_row9_col3, #T_0b4e1_row9_col4, #T_0b4e1_row9_col5, #T_0b4e1_row9_col6, #T_0b4e1_row9_col7, #T_0b4e1_row9_col8, #T_0b4e1_row10_col0, #T_0b4e1_row10_col1, #T_0b4e1_row10_col2, #T_0b4e1_row10_col3, #T_0b4e1_row10_col4, #T_0b4e1_row10_col5, #T_0b4e1_row10_col6, #T_0b4e1_row10_col7, #T_0b4e1_row10_col8, #T_0b4e1_row11_col0, #T_0b4e1_row11_col1, #T_0b4e1_row11_col2, #T_0b4e1_row11_col3, #T_0b4e1_row11_col4, #T_0b4e1_row11_col5, #T_0b4e1_row11_col6, #T_0b4e1_row11_col7, #T_0b4e1_row11_col8, #T_0b4e1_row12_col0, #T_0b4e1_row12_col1, #T_0b4e1_row12_col2, #T_0b4e1_row12_col3, #T_0b4e1_row12_col4, #T_0b4e1_row12_col5, #T_0b4e1_row12_col6, #T_0b4e1_row12_col7, #T_0b4e1_row12_col8, #T_0b4e1_row13_col0, #T_0b4e1_row13_col1, #T_0b4e1_row13_col2, #T_0b4e1_row13_col3, #T_0b4e1_row13_col4, #T_0b4e1_row13_col5, #T_0b4e1_row13_col6, #T_0b4e1_row13_col7, #T_0b4e1_row13_col8, #T_0b4e1_row14_col0, #T_0b4e1_row14_col1, #T_0b4e1_row14_col2, #T_0b4e1_row14_col3, #T_0b4e1_row14_col4, #T_0b4e1_row14_col5, #T_0b4e1_row14_col6, #T_0b4e1_row14_col7, #T_0b4e1_row14_col8, #T_0b4e1_row15_col0, #T_0b4e1_row15_col1, #T_0b4e1_row15_col2, #T_0b4e1_row15_col3, #T_0b4e1_row15_col4, #T_0b4e1_row15_col5, #T_0b4e1_row15_col6, #T_0b4e1_row15_col7, #T_0b4e1_row15_col8, #T_0b4e1_row16_col0, #T_0b4e1_row16_col1, #T_0b4e1_row16_col2, #T_0b4e1_row16_col3, #T_0b4e1_row16_col4, #T_0b4e1_row16_col5, #T_0b4e1_row16_col6, #T_0b4e1_row16_col7, #T_0b4e1_row16_col8, #T_0b4e1_row17_col0, #T_0b4e1_row17_col1, #T_0b4e1_row17_col2, #T_0b4e1_row17_col3, #T_0b4e1_row17_col4, #T_0b4e1_row17_col5, #T_0b4e1_row17_col6, #T_0b4e1_row17_col7, #T_0b4e1_row17_col8, #T_0b4e1_row18_col0, #T_0b4e1_row18_col1, #T_0b4e1_row18_col2, #T_0b4e1_row18_col3, #T_0b4e1_row18_col4, #T_0b4e1_row18_col5, #T_0b4e1_row18_col6, #T_0b4e1_row18_col7, #T_0b4e1_row18_col8, #T_0b4e1_row19_col0, #T_0b4e1_row19_col1, #T_0b4e1_row19_col2, #T_0b4e1_row19_col3, #T_0b4e1_row19_col4, #T_0b4e1_row19_col5, #T_0b4e1_row19_col6, #T_0b4e1_row19_col7, #T_0b4e1_row19_col8 { text-align: center; } </style>

	真实值	预测值	严重污染预测概率	中度污染预测概率	优预测概率	无预测概率	良预测概率	轻度污染预测概率	重度污染预测概率
1097	严重污染	重度污染	0.300962	0.053690	0.002110	0.002733	0.001727	0.005712	0.633066
2784	优	良	0.001050	0.005165	0.331282	0.001943	0.618779	0.041574	0.000207
2440	良	良	0.001645	0.022439	0.061227	0.001645	0.718252	0.193896	0.000895
1694	优	优	0.003246	0.005873	0.672221	0.005097	0.286800	0.025789	0.000974
2494	良	良	0.001708	0.006182	0.043863	0.000259	0.844592	0.101828	0.001568
2270	轻度污染	轻度污染	0.002498	0.370101	0.001685	0.001179	0.161089	0.424686	0.038762
3477	良	良	0.000512	0.032570	0.332566	0.001983	0.524911	0.104358	0.003101
937	良	中度污染	0.003020	0.459037	0.002037	0.000806	0.056424	0.465143	0.013532
495	中度污染	中度污染	0.014296	0.516219	0.002520	0.002577	0.042034	0.322502	0.099852
798	重度污染	严重污染	0.136483	0.412313	0.015048	0.007002	0.021073	0.056832	0.351248
3375	优	优	0.000160	0.001549	0.974973	0.001326	0.019990	0.001005	0.000997
1747	优	优	0.005040	0.002388	0.909634	0.000909	0.074456	0.005327	0.002246
1487	重度污染	重度污染	0.035081	0.346609	0.002323	0.001477	0.011112	0.045734	0.557664
969	中度污染	中度污染	0.009966	0.667408	0.000767	0.002853	0.006944	0.102986	0.209076
2883	中度污染	中度污染	0.008358	0.351996	0.007579	0.006394	0.056464	0.448937	0.120271
655	轻度污染	轻度污染	0.001423	0.113934	0.002714	0.000850	0.231265	0.637323	0.012491
229	轻度污染	中度污染	0.002663	0.525737	0.000823	0.000796	0.023640	0.428292	0.018049
1116	良	良	0.000173	0.002357	0.047882	0.000070	0.918638	0.030178	0.000702
840	中度污染	中度污染	0.014819	0.517901	0.001452	0.000666	0.015056	0.252036	0.198070
32	良	良	0.000178	0.012099	0.004606	0.000033	0.797791	0.183794	0.001499