def alpha_12(self):
"""
formula: (sign(delta(volume, 1))*(-1 * delta(close, 1)))
"""
return np.sign(
_utils.delta(self.volume, period=1) * -1 *
_utils.delta(self.close, period=1))
def alpha_24(self):
"""
formula: ((((delta((sum(close, 100) / 100), 100) / delay(close, 100)) < 0.05) or
((delta((sum(close, 100) / 100), 100) / delay(close, 100)) == 0.05)) ?
(-1* (close - ts_min(close, 100))) : (-1* delta(close, 3)))
"""
condition = _utils.delta(_utils.ts_ma(self.close, 100),
100) / _utils.delay(self.close, 100) <= 0.05
alpha = -1 * _utils.delta(self.close, 3)
alpha[condition] = -1 * (self.close - _utils.ts_min(self.close, 100))
return alpha
def alpha_7(self):
"""
formula for alpha7: ((adv20 < volume) ? ((-1*ts_rank(abs(delta(close, 7)), 60))
*sign(delta(close, 7))) : (-1*1))
careful, the self.close should be len>60, so that nan won't generated
"""
adv20 = _utils.ts_ma(self.volume, 20)
alpha = -1 * _utils.ts_rank(abs(_utils.delta(self.close, 7)),
60) * np.sign(_utils.delta(self.close, 7))
alpha[adv20 >= self.volume] = -1
return alpha
def alpha_17(self):
"""
formula: (((-1 * rank(ts_rank(close, 10))) *
rank(delta(delta(close, 1), 1)))
*rank(ts_rank((volume / adv20), 5)))
:return:
"""
adv20 = _utils.ts_ma(self.volume, 20)
return -1 * (
_utils.rank(_utils.ts_rank(self.close, 10)) * _utils.rank(
_utils.delta(_utils.delta(self.close, period=1), period=1)) *
_utils.rank(_utils.ts_rank(self.volume / adv20, window=5)))
def alpha_2(self):
"""
formula for alpha2: (-1*correlation(rank(delta(log(volume), 2)), rank(((close - open) / open)), 6))
:return: time series df of all stocks
"""
x1 = _utils.rank(_utils.delta(np.log(self.volume), period=2))
x2 = _utils.rank((self.close - self.open) / self.open)
return -1 * _utils.correlation(x1, x2, window=6)
def alpha_31(self):
"""
formula: ((rank(rank(rank(decay_linear((-1* rank(rank(delta(close, 10)))), 10)))) +
rank((-1* delta(close, 3)))) + sign(scale(correlation(adv20, low, 12))))
:return:
"""
adv20 = _utils.ts_ma(self.volume, 20)
df = _utils.correlation(adv20, self.low, 12).replace([-np.inf, np.inf],
0).fillna(value=0)
p1 = _utils.rank(
_utils.rank(
_utils.rank(
_utils.decay_linear_pn((-1 * _utils.rank(
_utils.rank(_utils.delta(self.close, 10)))), 10))))
p2 = _utils.rank((-1 * _utils.delta(self.close, 3)))
p3 = np.sign(_utils.scale(df))
return p1 + p2 + p3
def alpha_22(self):
"""
formula: (-1 * (delta(correlation(high, volume, 5), 5) * rank(stddev(close, 20))))
"""
df = _utils.correlation(self.high, self.volume, 5)
df = df.replace([-np.inf, np.inf], 0).fillna(value=0)
return -1 * _utils.delta(df, 5) * _utils.rank(
_utils.stddev(self.close, 20))
def alpha_19(self):
"""
formula: ((-1 * sign(((close - delay(close, 7)) + delta(close, 7)))) * (1 + rank((1 + sum(returns,250)))))
be careful with the time range, less than 250 will be nan
"""
return ((-1 * np.sign((self.close - _utils.delay(self.close, 7)) +
_utils.delta(self.close, 7))) *
(1 + _utils.rank(1 + _utils.ts_sum(self.returns, 250))))
def alpha_14(self):
"""
formula:((-1 * rank(delta(returns, 3))) * correlation(open, volume, 10))
:return: dataframe need a lot stock to prevent inf
"""
df = _utils.correlation(self.open, self.volume, window=10)
df = df.replace([-np.inf, np.inf], 0).fillna(method='backfill')
return -1 * _utils.rank(_utils.delta(self.returns, period=3)) * df
def alpha_42(self):
"""
formula: (ts_rank((volume / adv20), 20) * ts_rank((-1 * delta(close, 7)), 8))
:return:
"""
adv20 = _utils.ts_ma(self.volume, 20)
return _utils.ts_rank(self.volume / adv20, 20) * _utils.ts_rank(
(-1 * _utils.delta(self.close, 7)), 8)
def alpha_34(self):
"""
formula: rank(((1 - rank((stddev(returns, 2) / stddev(returns, 5)))) + (1 - rank(delta(close, 1)))))
:return:
"""
inner = _utils.stddev(self.returns, 2) / _utils.stddev(self.returns, 5)
inner = inner.replace([-np.inf, np.inf], 1).fillna(value=1)
return _utils.rank(2 - _utils.rank(inner) -
_utils.rank(_utils.delta(self.close, 1)))
def alpha_11(self):
"""
formula: ((rank(ts_max((vwap - close), 3)) + rank(ts_min((vwap - close), 3)))*rank(delta(volume, 3)))
:return:df backfilled
"""
df = (_utils.rank(_utils.ts_max(self.vwap - self.close, window=3)) +
_utils.rank(_utils.ts_min(self.vwap - self.close, window=3))
) * _utils.delta(self.volume, period=3)
return df.fillna(method='backfill')
def alpha_30(self):
"""
formula: (((1.0 - rank(((sign((close - delay(close, 1))) + sign((delay(close, 1) - delay(close, 2)))) +
sign((delay(close, 2) - delay(close, 3))))))* sum(volume, 5)) / sum(volume, 20))
"""
delta_close = _utils.delta(self.close, 1) # (close - delay(close, 1)
inner = np.sign(delta_close) + np.sign(_utils.delay(
delta_close, 1)) + np.sign(_utils.delay(delta_close, 2))
df = (1.0 - _utils.rank(inner)) * _utils.ts_sum(
self.volume, 5) / _utils.ts_sum(self.volume, 20)
return df.fillna(method='backfill')
def alpha_10(self):
"""
formula: rank(((0 < ts_min(delta(close, 1), 4)) ? delta(close, 1) : ((ts_max(delta(close, 1), 4) < 0) ?
delta(close, 1) : (-1*delta(close, 1)))))
"""
delta_close = _utils.delta(self.close, period=1)
condition_1 = _utils.ts_min(delta_close, window=4) > 0
condition_2 = _utils.ts_max(delta_close, window=4) < 0
alpha = -1 * delta_close
alpha[condition_1 | condition_2] = delta_close
return _utils.rank(alpha)
def alpha_23(self):
"""
formula: (((sum(high, 20) / 20) < high) ? (-1 * delta(high, 2)) : 0)
:return zeros dataframe, hat function
"""
condition = _utils.ts_ma(self.high, 20) < self.high
alpha = pd.DataFrame(np.zeros_like(self.close),
index=self.close.index,
columns=self.close.columns)
alpha[condition] = -1 * _utils.delta(self.high, 2).fillna(value=0)
return alpha
def alpha_9(self):
"""
formula: ((0 < ts_min(delta(close, 1), 5)) ? delta(close, 1) : ((ts_max(delta(close, 1), 5) < 0)
? delta(close, 1) : (-1*delta(close, 1))))
"""
delta_close = _utils.delta(self.close, period=1)
condition_1 = _utils.ts_min(delta_close, window=5) > 0
condition_2 = _utils.ts_max(delta_close, window=5) < 0
alpha = -1 * delta_close
alpha[condition_1 | condition_2] = delta_close
return alpha
def alpha_50(self):
"""
formula: ((((-1 * ts_min(low, 5)) + delay(ts_min(low, 5), 5)) * rank(((sum(returns, 240) -sum(returns, 20))
/ 220))) * ts_rank(volume, 5))
:return: long range data needed
"""
return ((
(-1 * _utils.delta(_utils.ts_min(self.low, 5), 5)) * _utils.rank(
((_utils.ts_sum(self.returns, 240) -
_utils.ts_sum(self.returns, 20)) / 220))) *
_utils.ts_rank(self.volume, 5))
def alpha_49(self):
"""
formula: (((((delay(close, 20) - delay(close, 10)) / 10) - ((delay(close, 10) - close) / 10)) < (-1 *0.05)) ?
1 : ((-1 * 1) * (close - delay(close, 1))))
:return:
"""
inner = ((
(_utils.delay(self.close, 20) - _utils.delay(self.close, 10)) / 10)
- ((_utils.delay(self.close, 10) - self.close) / 10))
alpha = (-1 * _utils.delta(self.close))
alpha[inner < -0.05] = 1
return alpha
def alpha_38(self):
"""
formula: ((-1 * rank((delta(close, 7) * (1 - rank(decay_linear_pn((volume / adv20), 9)))))) *
(1 +rank(sum(returns, 250))))
:return:
"""
adv20 = _utils.ts_ma(self.volume, 20)
return ((-1 * _utils.rank(
_utils.delta(self.close, 7) *
(1 - _utils.rank(_utils.decay_linear_pn(
(self.volume / adv20), 9))))) *
(1 + _utils.rank(_utils.ts_ma(self.returns, 250))))
def alpha_45(self):
"""
formula: (0.25 < (((delay(close, 20) - delay(close, 10)) / 10) - ((delay(close, 10) - close) / 10))) ?
(-1 * 1) : (((((delay(close, 20) - delay(close, 10)) / 10) - ((delay(close, 10) - close) / 10)) < 0) ?
1 :((-1 * 1) * (close - delay(close, 1)))))
:return dataframe， inner has lot of nan
"""
inner = ((_utils.delay(self.close, 20) - _utils.delay(self.close, 10)) / 10) - \
((_utils.delay(self.close, 10) - self.close) / 10)
alpha = (-1 * _utils.delta(self.close))
alpha[inner < 0] = 1
alpha[inner > 0.25] = -1
return alpha
def alpha_29(self):
"""
formula: (min(product(rank(rank(scale(log(sum(ts_min(rank(
rank((-1*rank(delta((close - 1), 5))))), 2), 1))))), 1), 5) +
ts_rank(delay((-1* returns), 6), 5))
"""
df = _utils.ts_min(
_utils.rank(
_utils.rank(
_utils.scale(
np.log(
_utils.ts_sum(
_utils.rank(
_utils.rank(-1 * _utils.rank(
_utils.delta(self.close - 1, 5)))),
2))))), 5) + _utils.ts_rank(
_utils.delay(-1 * self.returns, 6), 5)
return df.fillna(method='backfill')