-
Notifications
You must be signed in to change notification settings - Fork 86
/
eps.py
192 lines (143 loc) · 7.78 KB
/
eps.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
#!/usr/bin/env python
from regress import *
from loaddata import *
from util import *
from pandas.stats.moments import ewma
def wavg(group):
b = group['pbeta']
d = group['log_ret']
w = group['mkt_cap_y'] / 1e6
print "Mkt return: {} {}".format(group['gdate'], ((d * w).sum() / w.sum()))
res = b * ((d * w).sum() / w.sum())
return res
def calc_eps_daily(daily_df, horizon):
print "Caculating daily eps..."
result_df = filter_expandable(daily_df)
print "Calculating eps0..."
halflife = horizon / 2
# result_df['dk'] = np.exp( -1.0 * halflife * (result_df['gdate'] - result_df['last']).astype('timedelta64[D]').astype(int) )
# result_df['bret'] = result_df[['log_ret', 'pbeta', 'mkt_cap_y', 'gdate']].groupby('gdate').apply(wavg).reset_index(level=0)['pbeta']
# result_df['badjret'] = result_df['log_ret'] - result_df['bret']
# result_df['badj0_B'] = winsorize_by_date(result_df[ 'badjret' ])
# result_df['cum_ret'] = pd.rolling_sum(result_df['log_ret'], horizon)
result_df['std_diff'] = result_df['EPS_std'].unstack().diff().stack()
result_df.loc[ (result_df['std_diff'] <= 0) | (result_df['std_diff'].isnull()), 'EPS_diff_mean'] = 0
print "SEAN2"
print result_df.xs(testid, level=1)
result_df['eps0'] = result_df['EPS_diff_mean'] / result_df['EPS_median']
# print result_df.columns
# result_df['sum'] = result_df['EPS_median']
# result_df['det_diff'] = (result_df['sum'].diff())
# result_df['det_diff_sum'] = pd.rolling_sum( result_df['det_diff'], window=2)
# #result_df['det_diff_dk'] = ewma(result_df['det_diff'], halflife=horizon )
# result_df['eps0'] = result_df['det_diff']
# result_df['median'] = -1.0 * (result_df['median'] - 3)
# result_df['med_diff'] = result_df['median'].unstack().diff().stack()
# result_df['med_diff_dk'] = pd.rolling_sum( result_df['dk'] * result_df['med_diff'], window=horizon )
# result_df['eps0'] = (np.sign(result_df['med_diff_dk']) * np.sign(result_df['cum_ret'])).clip(lower=0) * result_df['med_diff_dk']
# demean = lambda x: (x - x.mean())
# indgroups = result_df[['eps0', 'gdate', 'ind1']].groupby(['gdate', 'ind1'], sort=True).transform(demean)
# result_df['eps0_ma'] = indgroups['eps0']
# result_df['eps0_ma'] = result_df['eps0_ma'] * (np.sign(result_df['eps0_ma']) * np.sign(result_df['cum_ret']))
result_df['eps0_ma'] = result_df['eps0']
for lag in range(1,horizon+1):
shift_df = result_df.unstack().shift(lag).stack()
result_df['eps'+str(lag)+'_ma'] = shift_df['eps0_ma']
return result_df
def eps_fits(daily_df, horizon, name, middate=None):
insample_daily_df = daily_df
if middate is not None:
insample_daily_df = daily_df[ daily_df.index.get_level_values('date') < middate ]
outsample_daily_df = daily_df[ daily_df.index.get_level_values('date') >= middate ]
print insample_daily_df['eps0_ma'].describe()
print outsample_daily_df['eps0_ma'].describe()
outsample_daily_df['eps'] = np.nan
fits_df = pd.DataFrame(columns=['horizon', 'coef', 'indep', 'tstat', 'nobs', 'stderr'])
for ii in range(1, horizon+1):
fitresults_df = regress_alpha(insample_daily_df, 'eps0_ma', ii, False, 'daily', False)
fits_df = fits_df.append(fitresults_df, ignore_index=True)
plot_fit(fits_df, "eps_daily_"+name+"_" + df_dates(insample_daily_df))
fits_df.set_index(keys=['indep', 'horizon'], inplace=True)
coef0 = fits_df.ix['eps0_ma'].ix[horizon].ix['coef']
print "Coef{}: {}".format(0, coef0)
outsample_daily_df[ 'eps0_ma_coef' ] = coef0
for lag in range(1,horizon):
coef = coef0 - fits_df.ix['eps0_ma'].ix[lag].ix['coef']
print "Coef{}: {}".format(lag, coef)
outsample_daily_df[ 'eps'+str(lag)+'_ma_coef' ] = coef
print "SEAN1"
print outsample_daily_df['eps0_ma'].describe()
outsample_daily_df[ 'eps' ] = outsample_daily_df['eps0_ma'].fillna(0) * outsample_daily_df['eps0_ma_coef']
print outsample_daily_df['eps'].describe()
for lag in range(1,horizon):
outsample_daily_df[ 'eps'] += outsample_daily_df['eps'+str(lag)+'_ma'].fillna(0) * outsample_daily_df['eps'+str(lag)+'_ma_coef']
print outsample_daily_df['eps'].describe()
return outsample_daily_df
def calc_eps_forecast(daily_df, horizon, middate):
daily_results_df = calc_eps_daily(daily_df, horizon)
forwards_df = calc_forward_returns(daily_df, horizon)
daily_results_df = pd.concat( [daily_results_df, forwards_df], axis=1)
#results = list()
# for sector_name in daily_results_df['sector_name'].dropna().unique():
# print "Running eps for sector {}".format(sector_name)
# sector_df = daily_results_df[ daily_results_df['sector_name'] == sector_name ]
# result_df = eps_fits(sector_df, horizon, sector_name, middate)
# results.append(result_df)
# result_df = pd.concat(results, verify_integrity=True)
# result_df = eps_fits(daily_results_df, horizon, "", middate)
# daily_results_df = daily_results_df[ daily_results_df['det_diff'] > 0]
# results = list()
# sector_name = 'Energy'
# print "Running eps for sector {}".format(sector_name)
# sector_df = daily_results_df[ daily_results_df['sector_name'] == sector_name ]
# res1 = eps_fits( sector_df[ sector_df['det_diff'] > 0 ], horizon, "energy_up", middate)
# # res2 = eps_fits( sector_df[ sector_df['det_diff'] < 0 ], horizon, "energy_dn", middate)
# results.append(res1)
# # results.append(res2)
# print "Running eps for not sector {}".format(sector_name)
# sector_df = daily_results_df[ daily_results_df['sector_name'] != sector_name ]
# res1 = eps_fits( sector_df[ sector_df['det_diff'] > 0 ], horizon, "rest_up", middate)
# # res2 = eps_fits( sector_df[ sector_df['det_diff'] < 0 ], horizon, "rest_dn", middate)
# results.append(res1)
# # results.append(res2)
# result_df = pd.concat(results, verify_integrity=True)
# res1 = eps_fits( daily_results_df[ daily_results_df['EPS_diff_mean'] > 0 ], horizon, "up", middate)
# res2 = eps_fits( daily_results_df[ daily_results_df['EPS_diff_mean'] < 0 ], horizon, "both", middate)
res2 = eps_fits( daily_results_df, horizon, "both", middate)
result_df = pd.concat([res2], verify_integrity=True)
return result_df
if __name__=="__main__":
parser = argparse.ArgumentParser(description='G')
parser.add_argument("--start",action="store",dest="start",default=None)
parser.add_argument("--end",action="store",dest="end",default=None)
parser.add_argument("--mid",action="store",dest="mid",default=None)
parser.add_argument("--lag",action="store",dest="lag",default=20)
# parser.add_argument("--horizon",action="store",dest="horizon",default=20)
args = parser.parse_args()
start = args.start
end = args.end
lookback = 30
horizon = int(args.lag)
pname = "./eps" + start + "." + end
start = dateparser.parse(start)
end = dateparser.parse(end)
middate = dateparser.parse(args.mid)
lag = int(args.lag)
loaded = False
try:
daily_df = pd.read_hdf(pname+"_daily.h5", 'table')
loaded = True
except:
print "Did not load cached data..."
if not loaded:
uni_df = get_uni(start, end, lookback)
BARRA_COLS = ['ind1', 'pbeta']
barra_df = load_barra(uni_df, start, end, BARRA_COLS)
PRICE_COLS = ['close']
price_df = load_prices(uni_df, start, end, PRICE_COLS)
daily_df = merge_barra_data(price_df, barra_df)
analyst_df = load_estimate_hist(price_df[['ticker']], start, end, "EPS")
daily_df = merge_daily_calcs(analyst_df, daily_df)
daily_df.to_hdf(pname+"_daily.h5", 'table', complib='zlib')
result_df = calc_eps_forecast(daily_df, horizon, middate)
dump_daily_alpha(result_df, 'eps')