forked from zll0622/Modeling_Analysis
45 lines
1.3 KiB
Python
45 lines
1.3 KiB
Python
import pandas as pd
|
||
import numpy as np
|
||
from sklearn.preprocessing import StandardScaler
|
||
|
||
def load_data(path):
|
||
"""加载CSV数据文件（支持UTF-8和GBK编码自动检测）"""
|
||
for enc in ["utf-8", "gbk", "latin-1"]:
|
||
try:
|
||
return pd.read_csv(path, encoding=enc)
|
||
except:
|
||
continue
|
||
return pd.read_csv(path)
|
||
|
||
def clean_data(df):
|
||
"""清洗数据：删除含缺失值的行，重置索引"""
|
||
df = df.dropna().reset_index(drop=True)
|
||
return df
|
||
|
||
def compute_corr_matrix(df):
|
||
"""计算相关系数矩阵（排除非数值列）"""
|
||
return df.select_dtypes(include=[np.number]).corr()
|
||
|
||
def detect_outliers(df, col):
|
||
"""检测异常值（超过99%分位数）"""
|
||
threshold = df[col].quantile(0.99)
|
||
return df[df[col] > threshold]
|
||
|
||
def normalize(df, cols):
|
||
"""标准化数据列"""
|
||
scaler = StandardScaler()
|
||
df[cols] = scaler.fit_transform(df[cols])
|
||
return df
|
||
|
||
def split_by_season(df):
|
||
"""按季节分割数据"""
|
||
seasons = {1: '冬', 2: '春', 3: '夏', 4: '秋'}
|
||
df['season'] = df['month'].apply(lambda m: seasons.get((m % 12 + 3) // 3, '未知'))
|
||
return df
|
||
|
||
def create_lag_features(df, col, lags):
|
||
"""创建滞后特征"""
|
||
for lag in lags:
|
||
df[f'{col}_lag{lag}'] = df[col].shift(lag)
|
||
return df
|