随机森林预测、重要性分析（Python实现）_随机森林预测环境因子重要性

from sklearn.model_selection import train_test_split
from sklearn.ensemble import RandomForestRegressor, RandomForestClassifier
from functools import reduce
import numpy as np
import pandas as pd
 
 
# 数据导入及基本信息定义
data = pd.read_excel('data2（Topsis评分评级）.xlsx')
data = data.drop(columns=['ID'])
 
prediction_set = data[data['MM'].isna()]
training_set = data.dropna()
 
features = ['XXX','XXX'] # 此处放表格列名
X = training_set[features]
y = training_set['MM']
minrmse = 1000
maxscore = 0
 
# 选取最优的random_state
'''
for randomstate in range(50, -1, -1):
    X_train, X_test, y_train, y_test = train_test_split(
        X, y, test_size=0.2, random_state=randomstate)
    model = RandomForestClassifier(random_state=randomstate)
    model.fit(X_train, y_train)
    y_pred = model.predict(X_test)
    score = accuracy_score(y_test, y_pred)
    rmse = np.sqrt(mean_squared_error(y_test, y_pred))
    if score > maxscore:
        maxscore = score
        print(maxscore, randomstate)
    X_pred = prediction_set[features]
    y_pred_prediction_set = model.predict(X_pred)
'''
# 第一种方法：随机森林分类器
X_train, X_test, y_train, y_test = train_test_split(X,
                                                    y,
                                                    test_size=0.2,
                                                    random_state=27)
 
model = RandomForestClassifier(random_state=27)
model.fit(X_train, y_train)
 
y_pred = model.predict(X_test)
X_pred = prediction_set[features]
y_pred_prediction_set = model.predict(X_pred)
y_pred_prediction_set = pd.DataFrame(y_pred_prediction_set)
y_pred_prediction_set.to_excel('Topsis评级预测-后20.xlsx')
 
# 第二种方法：随机森林回归器
X_train, X_test, y_train, y_test = train_test_split(X,
                                                    y,
                                                    test_size=0.2,
                                                    random_state=27)
 
model = RandomForestClassifier(random_state=27)
model.fit(X_train, y_train)
 
y_pred = model.predict(X_test)
X_pred = prediction_set[features]
y_pred_prediction_set = model.predict(X_pred)
y_pred_prediction_set = pd.DataFrame(y_pred_prediction_set)
y_pred_prediction_set.to_excel('Topsis评级预测-后20.xlsx')
 
 
# 重要性分析
importances = model.feature_importances_
importances_df = pd.DataFrame({
    'Feature': features,
    'Importance_behavior': importances
})
 
dfs = [importances]
df = reduce(lambda x, y: pd.merge(x, y, on="Feature", how="outer"), dfs)
df = pd.DataFrame(df)
df.to_excel('TopsisImportances-后20（8个特征重要性分析）.xlsx')

要注意的是，如果用分类器，y的取值需要是离散数值

如果用回归器，不要求是离散数据，但需要是数值
所以两种方法都要对目标列先进行数值化处理