Python中如何分析时间序列数据？

import pandas as pd
<h1>创建一个时间序列</h1><p>date_rng = pd.date_range(start='2023-01-01', end='2023-12-31', freq='D')
df = pd.DataFrame(date_rng, columns=['date'])
df['data'] = range(len(df))</p><h1>设置日期为索引</h1><p>df.set_index('date', inplace=True)</p><p>print(df.head())</p>

# 计算7天的移动平均
df['rolling_mean'] = df['data'].rolling(window=7).mean()
<p>print(df[['data', 'rolling_mean']].head(10))</p>

# 假设我们有一个每小时的温度数据
hourly_data = pd.read_csv('hourly_temperature.csv', index_col='datetime', parse_dates=True)
<h1>重采样为每天的平均温度</h1><p>daily_avg = hourly_data.resample('D').mean()</p><p>print(daily_avg.head())</p>

from statsmodels.tsa.arima.model import ARIMA
<h1>假设我们有一个时间序列数据</h1><p>ts = df['data']</p><h1>拟合ARIMA模型</h1><p>model = ARIMA(ts, order=(1,1,1))
results = model.fit()</p><h1>进行预测</h1><p>forecast = results.forecast(steps=30)</p><p>print(forecast)</p>

from statsmodels.tsa.seasonal import seasonal_decompose
<p>decomposition = seasonal_decompose(df['data'], model='additive', period=365)
trend = decomposition.trend
seasonal = decomposition.seasonal
residual = decomposition.resid</p><p>print(trend.head())
print(seasonal.head())
print(residual.head())</p>

from sklearn.model_selection import GridSearchCV
from statsmodels.tsa.arima.model import ARIMA
<h1>定义参数网格</h1><p>param_grid = {'order': [(p,d,q) for p in range(3) for d in range(2) for q in range(3)]}</p><h1>进行网格搜索</h1><p>grid_search = GridSearchCV(ARIMA(ts, order=(1,1,1)), param_grid, cv=5)
grid_search.fit()</p><p>print(grid_search.best<em>params</em>)</p>

import dask.dataframe as dd
<h1>假设我们有一个大规模的时间序列数据</h1><p>df = dd.read_csv('large_time_series.csv', parse_dates=['datetime'])</p><h1>进行重采样</h1><p>daily_avg = df.resample('D', on='datetime').mean().compute()</p><p>print(daily_avg.head())</p>

# 计算每周的销售总额
weekly_sales = sales_data.resample('W').sum()
<h1>注释解释每一步的作用</h1><h1>weekly_sales 包含每周的销售总额数据</h1><p>print(weekly_sales.head())</p>