calc_intraday_trend

gistfile1.txt

import numpy as np
from pandas.tseries.offsets import *
from collections import defaultdict
import statsmodels.api as sm
result_coefficient = defaultdict(lambda: defaultdict(dict))
result_tvalues = defaultdict(lambda: defaultdict(dict))
for x, intraday_data in dataset.items():
    intraday_data = intraday_data.set_index('time')
    days = intraday_data.resample('B').count().volume
    for day in days.index:
        if days[day] < 10:
            continue
        eod = day + DateOffset(hours=23, minutes=59, seconds=59)
        todays_data = intraday_data.loc[day:eod].close
        todays_data = todays_data.pct_change()
        todays_data = pd.DataFrame({'Y': todays_data,
                                    'X = Y-1': todays_data.shift(1), 
                                    'X = Y-2': todays_data.shift(2),
                                    'X = Y-3': todays_data.shift(3),
                                   }).dropna()
        for regressor in ['X = Y-1', 'X = Y-2', 'X = Y-3']:
            model = sm.OLS(todays_data['Y'], todays_data[regressor])
            model_result = model.fit()
            result_coefficient[x][(regressor, day)] = model_result.params[regressor]
            result_tvalues[x][(regressor, day)] = model_result.tvalues[regressor]