Chapter8.金融时间序列

最新推荐文章于 2024-03-12 11:19:57 发布

浮生若梦为欢几何？

最新推荐文章于 2024-03-12 11:19:57 发布

阅读量2.8k

点赞数

本文主要探讨了金融时间序列的数据特性，包括经济数据的概括型统计量、一段时间内的变化趋势和重抽样处理。此外，还介绍了移动统计值在技术分析中的应用，如移动平均，并结合实例进行了详细阐述。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

Chapter8.金融时间序列

import matplotlib as mpl
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns

mpl.rcParams['font.sans-serif'] = ['KaiTi']
mpl.rcParams['font.serif'] = ['KaiTi']
mpl.rcParams['axes.unicode_minus'] = False # 解决保存图像是负号'-'显示为方块的问题,或者转换负号为字符串
sns.set_style("darkgrid",{"font.sans-serif":['KaiTi', 'Arial']})

一、文章结构

1）经济数据
2）移动统计量，如：移动平均
3）相关性分析
4）高频数据
ps:本章节使用pandas 可以阅读《python for analysis》进一步了解

二、经济数据

#使用pandas读取数据更为方便

data = pd.read_csv(r"C:\Users\Administrator\Desktop\source\tr_eikon_eod_data.csv",index_col=0,parse_dates=True)
data             #index_col 指定索引列为第几列           parse_dates 是否将类似时间的文本数据，解析为 date数据

	AAPL.O	MSFT.O	INTC.O	AMZN.O	GS.N	SPY	.SPX	.VIX	EUR=	XAU=	GDX	GLD
Date
2010-01-01	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.4323	1096.35	NaN	NaN
2010-01-04	30.572827	30.950	20.88	133.90	173.08	113.33	1132.99	20.04	1.4411	1120.00	47.71	109.80
2010-01-05	30.625684	30.960	20.87	134.69	176.14	113.63	1136.52	19.35	1.4368	1118.65	48.17	109.70
2010-01-06	30.138541	30.770	20.80	132.25	174.26	113.71	1137.14	19.16	1.4412	1138.50	49.34	111.51
2010-01-07	30.082827	30.452	20.60	130.00	177.67	114.19	1141.69	19.06	1.4318	1131.90	49.10	110.82
...	...	...	...	...	...	...	...	...	...	...	...	...
2018-06-25	182.170000	98.390	50.71	1663.15	221.54	271.00	2717.07	17.33	1.1702	1265.00	22.01	119.89
2018-06-26	184.430000	99.080	49.67	1691.09	221.58	271.60	2723.06	15.92	1.1645	1258.64	21.95	119.26
2018-06-27	184.160000	97.540	48.76	1660.51	220.18	269.35	2699.63	17.91	1.1552	1251.62	21.81	118.58
2018-06-28	185.500000	98.630	49.25	1701.45	223.42	270.89	2716.31	16.85	1.1567	1247.88	21.93	118.22
2018-06-29	185.110000	98.610	49.71	1699.80	220.57	271.28	2718.37	16.09	1.1683	1252.25	22.31	118.65

2216 rows × 12 columns

data.head()

	AAPL.O	MSFT.O	INTC.O	AMZN.O	GS.N	SPY	.SPX	.VIX	EUR=	XAU=	GDX	GLD
Date
2010-01-01	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.4323	1096.35	NaN	NaN
2010-01-04	30.572827	30.950	20.88	133.90	173.08	113.33	1132.99	20.04	1.4411	1120.00	47.71	109.80
2010-01-05	30.625684	30.960	20.87	134.69	176.14	113.63	1136.52	19.35	1.4368	1118.65	48.17	109.70
2010-01-06	30.138541	30.770	20.80	132.25	174.26	113.71	1137.14	19.16	1.4412	1138.50	49.34	111.51
2010-01-07	30.082827	30.452	20.60	130.00	177.67	114.19	1141.69	19.06	1.4318	1131.90	49.10	110.82

data.plot(figsize=(10,12),subplots=True)        #横坐标自动按年划分，十分智能化

array([<matplotlib.axes._subplots.AxesSubplot object at 0x0F134390>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0F140690>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EF48AF0>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EF65D90>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EEB2FF0>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EECFCF0>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EEDDF50>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EF071D0>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EF07750>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EF78A70>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EFB7E70>,
       <matplotlib.axes._subplots.AxesSubplot object at 0x0EFD5B30>],
      dtype=object)

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-tmrZRX3H-1581857844288)(output_9_1.png)]

instruments = ['Apple Stock', 'Microsoft Stock',
'Intel Stock', 'Amazon Stock', 'Goldman Sachs Stock',
'SPDR S&P 500 ETF Trust', 'S&P 500 Index',
'VIX Volatility Index', 'EUR/USD Exchange Rate',
'Gold Price', 'VanEck Vectors Gold Miners ETF',
 'SPDR Gold Trust']

#将列名与实际名字一一对应，使用zip函数

for ric,name in zip(data.columns,instruments):
    print("{:8s}|{}".format(ric,name))            #foramt的正则格式：  8个长度的space

AAPL.O  |Apple Stock
MSFT.O  |Microsoft Stock
INTC.O  |Intel Stock
AMZN.O  |Amazon Stock
GS.N    |Goldman Sachs Stock
SPY     |SPDR S&P 500 ETF Trust
.SPX    |S&P 500 Index
.VIX    |VIX Volatility Index
EUR=    |EUR/USD Exchange Rate
XAU=    |Gold Price
GDX     |VanEck Vectors Gold Miners ETF
GLD     |SPDR Gold Trust

概括型统计量

data.info()      #数据大概信息

<class 'pandas.core.frame.DataFrame'>
DatetimeIndex: 2216 entries, 2010-01-01 to 2018-06-29
Data columns (total 12 columns):
AAPL.O    2138 non-null float64
MSFT.O    2138 non-null float64
INTC.O    2138 non-null float64
AMZN.O    2138 non-null float64
GS.N      2138 non-null float64
SPY       2138 non-null float64
.SPX      2138 non-null float64
.VIX      2138 non-null float64
EUR=      2216 non-null float64
XAU=      2211 non-null float64
GDX       2138 non-null float64
GLD       2138 non-null float64
dtypes: float64(12)
memory usage: 225.1 KB

data.describe()     #数据的统计信息

	AAPL.O	MSFT.O	INTC.O	AMZN.O	GS.N	SPY	.SPX	.VIX	EUR=	XAU=	GDX	GLD
count	2138.000000	2138.000000	2138.000000	2138.000000	2138.000000	2138.000000	2138.000000	2138.000000	2216.000000	2211.000000	2138.000000	2138.00000
mean	93.455973	44.561115	29.364192	480.461251	170.216221	180.323029	1802.713106	17.027133	1.248587	1349.014130	33.566525	130.08659
std	40.551559	19.527712	8.172160	372.307818	42.483935	48.190071	483.337146	5.883428	0.113711	188.745856	15.169564	18.78359
min	27.435687	23.010000	17.665000	108.610000	87.700000	102.200000	1022.580000	9.140000	1.038500	1051.360000	12.470000	100.50000
25%	60.294225	28.566875	22.510000	213.595000	146.607500	133.987500	1338.567500	13.072500	1.131500	1221.525000	22.142500	117.40250
50%	90.550624	39.660000	27.330000	322.065000	164.430000	186.320000	1863.080000	15.585000	1.270250	1292.610000	25.625000	124.00500
75%	117.235000	54.372500	34.710000	698.850000	192.132500	210.987500	2108.942500	19.067500	1.345400	1428.240000	48.342500	139.00250
max	193.980000	102.490000	57.080000	1750.080000	273.380000	286.580000	2872.870000	48.000000	1.482600	1898.990000	66.630000	184.59000

#aggregate,可以缩写为agg 用处比较广泛，参数可以是自定义函数

data.agg([np.min,np.max,np.mean])     #内置的统计函数

	AAPL.O	MSFT.O	INTC.O	AMZN.O	GS.N	SPY	.SPX	.VIX	EUR=	XAU=	GDX	GLD
amin	27.435687	23.010000	17.665000	108.610000	87.700000	102.200000	1022.580000	9.140000	1.038500	1051.36000	12.470000	100.50000
amax	193.980000	102.490000	57.080000	1750.080000	273.380000	286.580000	2872.870000	48.000000	1.482600	1898.99000	66.630000	184.59000
mean	93.455973	44.561115	29.364192	480.461251	170.216221	180.323029	1802.713106	17.027133	1.248587	1349.01413	33.566525	130.08659

#比如：我想自定义一个统计量，如极差
def my_jicha(arr):
    return arr.max() - arr.min()
data.agg(my_jicha)

AAPL.O     166.544313
MSFT.O      79.480000
INTC.O      39.415000
AMZN.O    1641.470000
GS.N       185.680000
SPY        184.380000
.SPX      1850.290000
.VIX        38.860000
EUR=         0.444100
XAU=       847.630000
GDX         54.160000
GLD         84.090000
dtype: float64

一段时间内变化

data.diff().mean()           #n阶差分,默认值为1     
#这个的实际含义：一段时间内，数据与前一数据相比产生的变化，对这个变化求平均；
#狭义理解一下，取绝对值后，数据越大，波动越大；若以股票为例，+代表涨，-代表跌

AAPL.O    0.064737
MSFT.O    0.031246
INTC.O    0.013540
AMZN.O    0.706608
GS.N      0.028224
SPY       0.072103
.SPX      0.732659
.VIX     -0.019583
EUR=     -0.000119
XAU=      0.041887
GDX      -0.015071
GLD      -0.003455
dtype: float64

#数值大小往往不具参考意义，我们更喜欢比率的变化
#所以，引入了pct_change  来表示n期的变化比率

bar_data = data.pct_change().mean()
bar_data

AAPL.O    0.000936
MSFT.O    0.000621
INTC.O    0.000499
AMZN.O    0.001332
GS.N      0.000242
SPY       0.000436
.SPX      0.000438
.VIX      0.002888
EUR=     -0.000075
XAU=      0.000109
GDX      -0.000072
GLD       0.000085
dtype: float64

bar_data.plot(kind="bar")       #个人觉得：pd.plot更为方便

<matplotlib.axes._subplots.AxesSubplot at 0x11f83750>

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-UmZxnf23-1581857844293)(output_23_1.png)]

#进行取对差分

rets = np.log(data/data.shift(1))  #由log性质可知，这是将数据取对之后，再进行差分

rets.cumsum().apply(np.exp).plot(figsize=(12,10))       #Q：实际意义是什么？

<matplotlib.axes._subplots.AxesSubplot at 0x13337a30>

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-EnIhSL1o-1581857844294)(output_26_1.png)]

重抽样

#改变时间间隔，比如：原时间序列时间间隔为日，改变为周；原间隔为周，改变为年

data.head()

	AAPL.O	MSFT.O	INTC.O	AMZN.O	GS.N	SPY	.SPX	.VIX	EUR=	XAU=	GDX	GLD
Date
2010-01-01	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.4323	1096.35	NaN	NaN
2010-01-04	30.572827	30.950	20.88	133.90	173.08	113.33	1132.99	20.04	1.4411	1120.00	47.71	109.80
2010-01-05	30.625684	30.960	20.87	134.69	176.14	113.63	1136.52	19.35	1.4368	1118.65	48.17	109.70
2010-01-06	30.138541	30.770	20.80	132.25	174.26	113.71	1137.14	19.16	1.4412	1138.50	49.34	111.51
2010-01-07	30.082827	30.452	20.60	130.00	177.67	114.19	1141.69	19.06	1.4318	1131.90	49.10	110.82

data.resample("3W",label="right").last()      #间隔为3周

	AAPL.O	MSFT.O	INTC.O	AMZN.O	GS.N	SPY	.SPX	.VIX	EUR=	XAU=	GDX	GLD
Date
2010-01-03	NaN	NaN	NaN	NaN	NaN	NaN	NaN	NaN	1.4323	1096.35	NaN	NaN
2010-01-24	28.249972	28.9600	19.91	121.43	154.12	109.210	1091.76	27.31	1.4137	1092.60	43.790	107.17
2010-02-14	28.625686	27.9300	20.43	119.66	153.93	108.040	1075.51	22.73	1.3615	1092.40	43.940	107.04
2010-03-07	31.278540	28.5875	20.79	128.91	167.18	114.250	1138.70	17.42	1.3621	1131.65	46.430	110.81
2010-03-28	32.985681	29.6600	22.24	135.06	172.87	116.582	1166.59	17.77	1.3410	1105.60	43.712	108.59
...	...	...	...	...	...	...	...	...	...	...	...	...
2018-04-15	174.730000	93.0800	51.86	1430.79	255.92	265.150	2656.30	17.41	1.2329	1345.01	22.890	127.45
2018-05-06	183.830000	95.1600	52.78	1580.95	234.94	266.020	2663.42	14.77	1.1958	1314.65	22.650	124.54
2018-05-27	188.580000	98.3600	55.44	1610.15	235.01	272.150	2721.33	13.22	1.1650	1301.17	22.310	123.21
2018-06-17	188.840000	100.1300	55.11	1715.97	231.92	277.130	2779.66	11.98	1.1607	1279.00	22.230	121.34
2018-07-08	185.110000	98.6100	49.71	1699.80	220.57	271.280	2718.37	16.09	1.1683	1252.25	22.310	118.65

149 rows × 12 columns

data.resample("1m",label="left",closed="left").last()  #间隔为1个月

	AAPL.O	MSFT.O	INTC.O	AMZN.O	GS.N	SPY	.SPX	.VIX	EUR=	XAU=	GDX	GLD
Date
2009-12-31	27.437544	28.180	19.40	125.410	148.72	107.3900	1073.87	24.62	1.3862	1081.05	40.72	105.960
2010-01-31	29.231399	28.670	20.53	118.400	156.35	110.7400	1104.49	19.50	1.3625	1116.10	43.89	109.430
2010-02-28	33.692109	29.770	22.34	136.579	171.38	117.4000	1173.27	17.13	1.3411	1102.60	43.86	107.970
2010-03-31	38.377104	31.003	23.49	141.730	160.24	120.8575	1206.78	18.44	1.3224	1166.10	50.31	114.280
2010-04-30	36.697106	25.800	21.42	125.460	144.26	109.3690	1089.41	32.07	1.2267	1213.81	49.86	118.881
...	...	...	...	...	...	...	...	...	...	...	...	...
2018-01-31	178.390000	94.200	49.91	1511.980	267.93	274.4300	2744.28	18.59	1.2231	1318.14	21.49	125.130
2018-02-28	167.780000	91.270	52.08	1447.340	251.86	263.1500	2640.87	19.97	1.2321	1324.00	21.98	125.790
2018-03-31	162.320000	95.820	52.73	1572.620	239.80	266.5600	2669.91	15.41	1.2128	1321.89	22.73	125.500
2018-04-30	187.500000	98.950	55.68	1624.890	229.16	272.6100	2724.01	14.94	1.1661	1301.03	22.44	123.370
2018-05-31	185.110000	98.610	49.71	1699.800	220.57	271.2800	2718.37	16.09	1.1683	1252.25	22.31	118.650

102 rows × 12 columns

rets.cumsum().apply(np.exp).resample("1m",label="right").last().plot()

<matplotlib.axes._subplots.AxesSubplot at 0x1141fb10>

[外链图片转存失败,源站可能有防盗链机制,建议将图片保存下来直接上传(img-XiWauNvR-1581857844296)(output_32_1.png)]

三、移动统计值

data = pd.DataFrame(data['AAPL.O']).dropna()

data.head()

	AAPL.O
Date
2010-01-04	30.572827
2010-01-05	30.625684
2010-01-06	30.138541
2010-01-07	30.082827
2010-01-08	30.282827

综述

window = 20   #20日的移动

data["min"] = data['AAPL.O'].rolling(window=window).min()
data["median"] = data['AAPL.O'].rolling(window=window).median()
data["max"] = data['AAPL.O'].rolling(window=window).max()

data["std"] = data['AAPL.O'].rolling(window=window).std()
data["ewma"] = data['AAPL.O'].ewm(halflife=0.5,min_periods=window).mean()  #加权指数平滑，里面的参数不明白！！！！

data

	AAPL.O	min	std	ewma	median	max
Date
2010-01-04	30.572827	NaN	NaN	NaN	NaN	NaN
2010-01-05	30.625684	NaN	NaN	NaN	NaN	NaN
2010-01-06	30.138541	NaN	NaN	NaN	NaN	NaN
2010-01-07	30.082827	NaN	NaN	NaN	NaN	NaN
2010-01-08	30.282827	NaN	NaN	NaN	NaN	NaN
...	...	...	...	...	...	...
2018-06-25	182.170000	182.17	3.262549	182.907462	189.54	193.98
2018-06-26	184.430000	182.17	3.423931	184.049366	189.54	193.98
2018-06-27	184.160000	182.17	3.580486	184.132341	189.54	193.98
2018-06-28	185.500000	182.17	3.633386	185.158085	189.54	193.98
2018-06-29	185.110000	182.17	3.706497	185.122021	188.79	193.98

2138 rows × 6 columns

#选用数据进行作图可视化

%matplotlib notebook

fig1 = plt.figure()

<IPython.core.display.Javascript object>

data[["AAPL.O"]].plot()

<IPython.core.display.Javascript object>

<matplotlib.axes._subplots.AxesSubplot at 0x14e28e10>

data[["min","max","median"]].plot()

<IPython.core.display.Javascript object>

<matplotlib.axes._subplots.AxesSubplot at 0x14dd0e50>

一个技术分析的例子

#利用短期MA线 和 长期MA线 改进策略
#当短期MA 向上突破长期线 ： 可考虑买入
#当短期MA 向下突破长期线： 可考虑卖出

data["sma1"] = data["AAPL.O"].rolling(42).mean()
data["sma2"] = data["AAPL.O"].rolling(252).mean()
data[["AAPL.O","sma1","sma2"]].plot(style=["b-","g--","r--"])

<IPython.core.display.Javascript object>

<img src="data:image/png;base64,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