机器学习基础算法13-回归实例-时间序列分析ARIMA

模型介绍

https://blog.csdn.net/u012735708/article/details/82460962

实例介绍

给定某航班的乘客变化数据，使用ARIMA计算预测模型

代码

import pandas as pd
import numpy as np
from statsmodels.tsa.arima_model import ARIMA
import matplotlib as mpl
import matplotlib.pyplot as plt
import warnings
import matplotlib.patches as mpatches
from statsmodels.tools.sm_exceptions import HessianInversionWarning


def extend(a, b):
    return 1.05*a-0.05*b, 1.05*b-0.05*a


def date_parser(date):
    # 转换时间索引
    return pd.datetime.strptime(date, '%Y-%m')


if __name__ == '__main__':

    warnings.filterwarnings(action='ignore', category=HessianInversionWarning)
    # 设置显示宽度
    pd.set_option('display.width', 100)
    # 全部打印，参数suppress表示是否用科学计数法来表示浮点数
    np.set_printoptions(linewidth=100, suppress=True)
    # 参数parse_dates为把某一列解析为时间索引；date_parser返回一个包含指定日的详细信息的关联数组
    data = pd.read_csv('AirPassengers.csv', header=0, parse_dates=['Month'], date_parser=date_parser, index_col=['Month'])
    # 对列索引重新命名
    data.rename(columns={'#Passengers': 'Passengers'}, inplace=True)
    print(data.dtypes)

    mpl.rcParams['font.sans-serif'] = [u'SimHei']
    mpl.rcParams['axes.unicode_minus'] = False

    x = data['Passengers'].astype(np.float)
    # 对数计算-默认以e为底
    x = np.log(x)
    print(x.head(10))

    show = 'prime'   # 'diff', 'ma', 'prime'
    d = 1
    diff = x - x.shift(periods=d)
    ma = x.rolling(window=12).mean()
    xma = x - ma

    # 时间序列分析-ARIMA
    p = 2
    q = 2
    model = ARIMA(endog=x, order=(p, d, q))     # 自回归函数p,差分d,移动平均数q
    arima = model.fit(disp=-1)                  # disp<0:不输出过程
    prediction = arima.fittedvalues
    print(type(prediction))
    y = prediction.cumsum() + x[0]
    mse = ((x - y)**2).mean()
    rmse = np.sqrt(mse)

    plt.figure(facecolor='w')

    if show == 'diff':
        plt.plot(x, 'r-', lw=2, label=u'原始数据')
        plt.plot(diff, 'g-', lw=2, label=u'%d阶差分' % d)
        #plt.plot(prediction, 'r-', lw=2, label=u'预测数据')
        title = u'乘客人数变化曲线 - 取对数'
    elif show == 'ma':
        #plt.plot(x, 'r-', lw=2, label=u'原始数据')
        #plt.plot(ma, 'g-', lw=2, label=u'滑动平均数据')
        plt.plot(xma, 'g-', lw=2, label=u'ln原始数据 - ln滑动平均数据')
        plt.plot(prediction, 'r-', lw=2, label=u'预测数据')
        title = u'滑动平均值与MA预测值'
    else:
        plt.plot(x, 'r-', lw=2, label=u'原始数据')
        plt.plot(y, 'g-', lw=2, label=u'预测数据')
        title = u'对数乘客人数与预测值(AR=%d, d=%d, MA=%d)：RMSE=%.4f' % (p, d, q, rmse)

    plt.legend(loc='upper left')
    plt.grid(b=True, ls=':')
    plt.title(title, fontsize=18)
    plt.tight_layout(2)
    plt.savefig('%s.png' % title)
    plt.show()

实验结果

运行结果：

Passengers    int64
dtype: object
Month
1949-01-01    4.718499
1949-02-01    4.770685
1949-03-01    4.882802
1949-04-01    4.859812
1949-05-01    4.795791
1949-06-01    4.905275
1949-07-01    4.997212
1949-08-01    4.997212
1949-09-01    4.912655
1949-10-01    4.779123
Name: Passengers, dtype: float64

show = 'prime’

show = 'ma’

show = 'diff’