Python实现Pearson's r检验：计算两组变量的相关系数与p值

完整代码实现 #

# Import libraries
import pandas as pd
from scipy import stats

# Load DataFrame
df = pd.read_csv('https://tf-assets-prod.s3.amazonaws.com/tf-curric/data-analytics-bootcamp/medicalcosts.csv')

# Use the 'pearsonr' correlation test to check for correlation between `age` and `charges` and print the results
ac, ap = stats.pearsonr(df['age'], df['charges'])
print(f"年龄与医疗费用的Pearson相关系数: {ac:.4f}, p值: {ap:.4f}")

# Use the 'pearsonr' correlation test to check for correlation between `bmi` and `charges` and print the results
bc, bp = stats.pearsonr(df['bmi'], df['charges'])
print(f"BMI与医疗费用的Pearson相关系数: {bc:.4f}, p值: {bp:.4f}")

关键代码说明 #

• stats.pearsonr(x, y)：Scipy库中计算Pearson相关系数与p值的专用函数，输入两个等长数值数组后，返回**(相关系数, p值)**元组，可直接通过Python多变量赋值特性分配给目标变量。
• df['age']/df['bmi']/df['charges']：从数据集中提取对应列的数值数组，作为相关性分析的输入数据源。
• 格式化输出：使用f-string将结果保留4位小数，提升输出可读性。

内容的提问来源于stack exchange，提问作者Sarah Wallace