QUANTAXIS ANALYSIS 行情分析研究

• 1  QUANTAXIS ANALYSIS 行情分析/研究
  • 1.1  开盘价序列
  • 1.2  收盘价序列
  • 1.3  最高价序列
  • 1.4  最低价序列
  • 1.5  量
  • 1.6  同vol
  • 1.7  日期
  • 1.8  时间
  • 1.9  索引
  • 1.10  平均价(O+H+L+C)/4
  • 1.11  price 的平均数
  • 1.12  price 的最大值
  • 1.13  price 的最小值
  • 1.14  price 的平均绝对偏差
  • 1.15  price 的众数(没啥用)
  • 1.16  price 的一阶差分
  • 1.17  price 的方差
  • 1.18  price 的样本方差
  • 1.19  price 的标准差
  • 1.20  price 的样本标准差
  • 1.21  price 的调和平均数
  • 1.22  price 的振幅[极差]
  • 1.23  price 的峰度 (4阶中心距)
  • 1.24  price 的偏度 (3阶中心距)
  • 1.25  price 的百分比变化序列
• 2  函数指标计算
• 3  指标计算, 和DataStruct用法一致

QUANTAXIS ANALYSIS 行情分析/研究

• 主要是针对行情的各种统计学特征/指标等分析
• 支持QA_DataStruct_系列的add_func()功能
• 接收DataFrame形式的行情以及QUANTAXIS.QADATA格式的行情

import QUANTAXIS as QA

you are using non-interactive mdoel quantaxis

class QAAnalysis_stock():
    """
    行情分析器
    计算所有的指标
    """

    def __init__(self, dataStruct, *args, **kwargs):
        try:
            # 如果是QA_Data_系列
            self.data = dataStruct.data
            self._data = dataStruct
        except AttributeError:
            # 如果是dataframe
            self.data = dataStruct

        # self.data=DataSturct.data

    def __repr__(self):
        return '< QAAnalysis_Stock >'

    def __call__(self):
        return self.data

    # 使用property进行懒运算

    @property
    def open(self):
        return self.data['open']

    @property
    def high(self):
        return self.data['high']

    @property
    def low(self):
        return self.data['low']

    @property
    def close(self):
        return self.data['close']

    @property
    def vol(self):
        if 'volume' in self.data.columns:
            return self.data['volume']
        else:
            return self.data['vol']

    @property
    def volume(self):
        if 'volume' in self.data.columns:
            return self.data['volume']
        else:
            return self.data['vol']

    @property
    def date(self):

        return self.data.index.levels[self.data.index.names.index(
            'date')] if 'date' in self.data.index.names else self.data['date']

    @property
    def datetime(self):

        return self.data.index.levels[self.data.index.names.index(
            'datetime')] if 'datetime' in self.data.index.names else self.data.index.levels[self.data.index.names.index(
                'date')]

    @property
    def index(self):
        return self.data.index

    # 均价
    @property
    def price(self):
        return 0.25 * (self.open + self.close + self.high + self.low)

    @property
    def max(self):
        return self.price.max()

    @property
    def min(self):
        return self.price.min()

    @property
    def mean(self):
        return self.price.mean()
    # 一阶差分序列

    @property
    def price_diff(self):
        return self.price.diff(1)

    # 样本方差(无偏估计) population variance
    @property
    def pvariance(self):
        return statistics.pvariance(self.price)

    # 方差
    @property
    def variance(self):

        return statistics.variance(self.price)
    # 标准差

    @property
    def day_pct_change(self):
        return (self.open - self.close) / self.open

    @property
    def stdev(self):

        return statistics.stdev(self.price)
    # 样本标准差

    @property
    def pstdev(self):
        return statistics.pstdev(self.price)

    # 调和平均数
    @property
    def mean_harmonic(self):
        return statistics.harmonic_mean(self.price)

    # 众数
    @property
    def mode(self):
        return statistics.mode(self.price)

    # 波动率

    # 振幅
    @property
    def amplitude(self):
        return self.max - self.min
    # 偏度 Skewness

    @property
    def skewnewss(self):
        return self.price.skew()
    # 峰度Kurtosis

    @property
    def kurtosis(self):
        return self.price.kurt()
    # 百分数变化

    @property
    def pct_change(self):
        return self.price.pct_change()

    # 平均绝对偏差
    @property
    def mad(self):
        return self.price.mad()

    # 函数 指标计算
    @lru_cache()
    def add_func(self, func, *arg, **kwargs):
        return func(self.data, *arg, **kwargs)

data=QA.QA_fetch_stock_day_adv('300872','2020-01-01','2021-09-15') #[可选to_qfq(),to_hfq()]
s=QA.QAAnalysis_stock(data)

开盘价序列

    @property
    def open(self):
        return self.data['open']

s.open 

date        code  
2020-08-24  300872    46.69
2020-08-25  300872    66.15
2020-08-26  300872    78.50
2020-08-27  300872    73.01
2020-08-28  300872    58.00
                      ...  
2021-09-08  300872    32.71
2021-09-09  300872    32.98
2021-09-10  300872    32.28
2021-09-13  300872    32.35
2021-09-14  300872    32.20
Name: open, Length: 260, dtype: float64

收盘价序列

    @property
    def close(self):
        return self.data['close']

s.close

date        code  
2020-08-24  300872    76.53
2020-08-25  300872    82.10
2020-08-26  300872    70.99
2020-08-27  300872    60.00
2020-08-28  300872    52.15
                      ...  
2021-09-08  300872    33.15
2021-09-09  300872    32.45
2021-09-10  300872    32.54
2021-09-13  300872    31.92
2021-09-14  300872    31.77
Name: close, Length: 260, dtype: float64

最高价序列

    @property
    def high(self):
        return self.data['high']

s.high

date        code  
2020-08-24  300872    98.00
2020-08-25  300872    90.77
2020-08-26  300872    82.00
2020-08-27  300872    74.00
2020-08-28  300872    58.85
                      ...  
2021-09-08  300872    33.29
2021-09-09  300872    33.11
2021-09-10  300872    32.98
2021-09-13  300872    32.57
2021-09-14  300872    32.67
Name: high, Length: 260, dtype: float64

最低价序列

    @property
    def low(self):
        return self.data['low']

s.low 

date        code  
2020-08-24  300872    46.69
2020-08-25  300872    66.10
2020-08-26  300872    69.98
2020-08-27  300872    58.97
2020-08-28  300872    52.01
                      ...  
2021-09-08  300872    32.58
2021-09-09  300872    32.42
2021-09-10  300872    32.00
2021-09-13  300872    31.92
2021-09-14  300872    31.64
Name: low, Length: 260, dtype: float64

量

    @property
    def vol(self):
        if 'volume' in self.data.columns:
            return self.data['volume']
        else:
            return self.data['vol']

s.vol  

date        code  
2020-08-24  300872    337942.0
2020-08-25  300872    335380.0
2020-08-26  300872    216283.0
2020-08-27  300872    198959.0
2020-08-28  300872    164599.0
                        ...   
2021-09-08  300872     33459.0
2021-09-09  300872     26253.0
2021-09-10  300872     29084.0
2021-09-13  300872     22417.0
2021-09-14  300872     26838.0
Name: volume, Length: 260, dtype: float64

同vol

s.volume

date        code  
2020-08-24  300872    337942.0
2020-08-25  300872    335380.0
2020-08-26  300872    216283.0
2020-08-27  300872    198959.0
2020-08-28  300872    164599.0
                        ...   
2021-09-08  300872     33459.0
2021-09-09  300872     26253.0
2021-09-10  300872     29084.0
2021-09-13  300872     22417.0
2021-09-14  300872     26838.0
Name: volume, Length: 260, dtype: float64

日期

    @property
    def datetime(self):

        return self.data.index.levels[self.data.index.names.index(
            'datetime')] if 'datetime' in self.data.index.names else self.data.index.levels[self.data.index.names.index('date')]

s.date

DatetimeIndex(['2020-08-24', '2020-08-25', '2020-08-26', '2020-08-27',
               '2020-08-28', '2020-08-31', '2020-09-01', '2020-09-02',
               '2020-09-03', '2020-09-04',
               ...
               '2021-09-01', '2021-09-02', '2021-09-03', '2021-09-06',
               '2021-09-07', '2021-09-08', '2021-09-09', '2021-09-10',
               '2021-09-13', '2021-09-14'],
              dtype='datetime64[ns]', name='date', length=260, freq=None)

时间

s.datetime

DatetimeIndex(['2020-08-24', '2020-08-25', '2020-08-26', '2020-08-27',
               '2020-08-28', '2020-08-31', '2020-09-01', '2020-09-02',
               '2020-09-03', '2020-09-04',
               ...
               '2021-09-01', '2021-09-02', '2021-09-03', '2021-09-06',
               '2021-09-07', '2021-09-08', '2021-09-09', '2021-09-10',
               '2021-09-13', '2021-09-14'],
              dtype='datetime64[ns]', name='date', length=260, freq=None)

索引

    @property
    def index(self):
        return self.data.index

s.index

MultiIndex([('2020-08-24', '300872'),
            ('2020-08-25', '300872'),
            ('2020-08-26', '300872'),
            ('2020-08-27', '300872'),
            ('2020-08-28', '300872'),
            ('2020-08-31', '300872'),
            ('2020-09-01', '300872'),
            ('2020-09-02', '300872'),
            ('2020-09-03', '300872'),
            ('2020-09-04', '300872'),
            ...
            ('2021-09-01', '300872'),
            ('2021-09-02', '300872'),
            ('2021-09-03', '300872'),
            ('2021-09-06', '300872'),
            ('2021-09-07', '300872'),
            ('2021-09-08', '300872'),
            ('2021-09-09', '300872'),
            ('2021-09-10', '300872'),
            ('2021-09-13', '300872'),
            ('2021-09-14', '300872')],
           names=['date', 'code'], length=260)

平均价(O+H+L+C)/4

    # 均价
    @property
    def price(self):
        return 0.25 * (self.open + self.close + self.high + self.low)

s.price

date        code  
2020-08-24  300872    66.9775
2020-08-25  300872    76.2800
2020-08-26  300872    75.3675
2020-08-27  300872    66.4950
2020-08-28  300872    55.2525
                       ...   
2021-09-08  300872    32.9325
2021-09-09  300872    32.7400
2021-09-10  300872    32.4500
2021-09-13  300872    32.1900
2021-09-14  300872    32.0700
Length: 260, dtype: float64

price 的平均数

    @property
    def mean(self):
        return self.price.mean()

s.mean

39.904500000000006

price 的最大值

    @property
    def max(self):
        return self.price.max()

s.max

76.28

price 的最小值

    @property
    def min(self):
        return self.price.min()

s.min

29.909999999999997

price 的平均绝对偏差

• Mean Absolute Deviation
• 平均绝对离差(mean absolute deviation)简称“平均离差”
• 当总体的单位数为 N 时，有变量 X1，X2，X3，……，XN一1，XN
• 各项变量与总体平均数之差叫离差
• 平均绝对离差定义：为各数据与平均值的离差的绝对值的平均数

    # 平均绝对偏差
    @property
    def mad(self):
        return self.price.mad()

s.mad

6.63717307692308

price 的众数(没啥用)

    # 众数
    @property
    def mode(self):
        return statistics.mode(self.price)

s.mode

66.97749999999999

price 的一阶差分

    # 一阶差分序列

    @property
    def price_diff(self):
        return self.price.diff(1)

s.price_diff

date        code  
2020-08-24  300872        NaN
2020-08-25  300872     9.3025
2020-08-26  300872    -0.9125
2020-08-27  300872    -8.8725
2020-08-28  300872   -11.2425
                       ...   
2021-09-08  300872     0.0600
2021-09-09  300872    -0.1925
2021-09-10  300872    -0.2900
2021-09-13  300872    -0.2600
2021-09-14  300872    -0.1200
Length: 260, dtype: float64

price 的方差

    # 方差
    @property
    def variance(self):
        return statistics.variance(self.price)

s.variance 

69.37130583011583

price 的样本方差

    # 样本方差(无偏估计) population variance
    @property
    def pvariance(self):
        return statistics.pvariance(self.price)

s.pvariance

69.10449311538461

price 的标准差

    # 标准差
    @property
    def stdev(self):
        return statistics.stdev(self.price)

s.stdev

8.328943860425271

price 的样本标准差

    # 样本标准差
    @property
    def pstdev(self):
        return statistics.pstdev(self.price)

s.pstdev

8.312911229851105

price 的调和平均数

    # 调和平均数
    @property
    def mean_harmonic(self):
        return statistics.harmonic_mean(self.price)

s.mean_harmonic

38.493185546318685

price 的振幅[极差]

    # 振幅
    @property
    def amplitude(self):
        return self.max - self.min

s.amplitude

46.370000000000005

price 的峰度 (4阶中心距)

    # 偏度 Skewness

    @property
    def skewnewss(self):
        return self.price.skew()

s.skewnewss

1.4109494377363532

price 的偏度 (3阶中心距)

    # 峰度Kurtosis
    @property
    def kurtosis(self):
        return self.price.kurt()

s.kurtosis

2.156590388276748

price 的百分比变化序列

    # 百分数变化
    @property
    def pct_change(self):
        return self.price.pct_change()

s.pct_change 

date        code  
2020-08-24  300872         NaN
2020-08-25  300872    0.138890
2020-08-26  300872   -0.011963
2020-08-27  300872   -0.117723
2020-08-28  300872   -0.169073
                        ...   
2021-09-08  300872    0.001825
2021-09-09  300872   -0.005845
2021-09-10  300872   -0.008858
2021-09-13  300872   -0.008012
2021-09-14  300872   -0.003728
Length: 260, dtype: float64

函数指标计算

    @lru_cache()
    def add_func(self, func, *arg, **kwargs):
        return func(self.data, *arg, **kwargs)

指标计算, 和DataStruct用法一致

s.add_func(QA.QA_indicator_CCI)

		CCI	a	b
date	code
2020-08-24	300872	NaN	100	-100
2020-08-25	300872	NaN	100	-100
2020-08-26	300872	NaN	100	-100
2020-08-27	300872	NaN	100	-100
2020-08-28	300872	NaN	100	-100
...	...	...	...	...
2021-09-08	300872	167.881484	100	-100
2021-09-09	300872	86.993291	100	-100
2021-09-10	300872	44.500035	100	-100
2021-09-13	300872	-37.419768	100	-100
2021-09-14	300872	-60.101682	100	-100

260 rows × 3 columns