mbp = (2 * sbp_and_dbp_mat[:, 0] + sbp_and_dbp_mat[:, 1]) / 3
return mbp
# return np.concatenate((sbp_and_dbp_mat, mbp), axis=1)
if __name__ == "__main__":
# a = [(0.098, 'RBW10'), (0.095, 'RBW25'), (0.089, 'RBW33'), (0.087, 'kte_delta'), (0.076, 'RBW50'), (0.069, 'pwtt_mean'), (0.067, 'RBW66'), (0.064, 'kte_skew'), (0.058, 'RBW75'), (0.058, 'DBW50'), (0.057, 'SLP2'), (0.057, 'SLP1'), (0.056, 'h_miu'), (0.055, 'kte_iqr'), (0.055, 'PRT'), (0.055, 'DBW33'), (0.053, 'DBW25'), (0.051, 'DBW10'), (0.049, 'DBW66'), (0.042, 'KVAL'), (0.042, 'DBW75'), (0.034, 'hr_delta'), (0.028, 'hr_miu'), (0.028, 'PH'), (0.027, 'ppg_fed_ar_5'), (0.027, 'RBAr'), (0.027, 'AmBE'), (0.026, 'kte_miu'), (0.026, 'PWA'), (0.021, 'DfAmBE'), (0.017, 'hr_skew'), (0.014, 'hr_iqr'), (0.013, 'ppg_fed_ar_4'), (0.009, 'ppg_fed_ar_1'), (0.009, 'h_delta'), (0.008, 'ppg_fed_ar_2'), (0.008, 'loge_delta'), (0.006, 'loge_iqr'), (0.005, 'h_iqr'), (0.002, 'ppg_fed_ar_3'), (0.001, 'h_skew'), (-0.0, 'loge_ar_5'), (-0.0, 'loge_ar_4'), (-0.0, 'loge_ar_3'), (-0.0, 'loge_ar_2'), (0.0, 'loge_ar_1')]
# b = [(0.086, 'kte_delta'), (0.064, 'hr_miu'), (0.055, 'kte_iqr'), (0.054, 'RBW75'), (0.052, 'RBW66'), (0.049, 'RBW50'), (0.045, 'hr_delta'), (0.043, 'RBW33'), (0.037, 'RBW25'), (0.035, 'PWA'), (0.035, 'DBW66'), (0.034, 'kte_miu'), (0.034, 'RBW10'), (0.034, 'DBW50'), (0.033, 'AmBE'), (0.032, 'pwtt_mean'), (0.032, 'DfAmBE'), (0.028, 'RBAr'), (0.028, 'PH'), (0.028, 'DBW75'), (0.026, 'DBW33'), (0.026, 'DBW25'), (0.026, 'DBW10'), (0.024, 'ppg_fed_ar_2'), (0.023, 'ppg_fed_ar_1'), (0.022, 'kte_skew'), (0.022, 'KVAL'), (0.021, 'SLP1'), (0.017, 'hr_skew'), (0.017, 'PRT'), (0.015, 'SLP2'), (0.014, 'hr_iqr'), (0.013, 'ppg_fed_ar_5'), (0.012, 'h_miu'), (0.01, 'ppg_fed_ar_3'), (0.009, 'loge_delta'), (0.009, 'h_iqr'), (0.007, 'loge_iqr'), (0.007, 'h_delta'), (0.003, 'ppg_fed_ar_4'), (0.002, 'loge_ar_1'), (0.002, 'h_skew'), (0.001, 'loge_ar_2'), (0.0, 'loge_ar_5'), (0.0, 'loge_ar_4'), (-0.0, 'loge_ar_3')]
# disp_map(a)
# print('*****************')
# disp_map(b)
# exit()
fh = FileHelper()
all_csv_names = fh.get_all_csv_names()
full_set_arr = []
full_set_res = []
for csv_file_name in all_csv_names:
is_valid, arr, res = fh.read_file(csv_file_name)
if not is_valid:
continue
full_set_arr = np.concatenate((full_set_arr, arr.tolist()), 0) if len(full_set_arr) > 0 else arr.tolist()
res = get_mean_bp(res)
full_set_res = np.concatenate((full_set_res, res.tolist()), 0) if len(full_set_res) > 0 else res.tolist()
# print '************' + fh.colsRes[0] + '***************'
# print rank_features(full_set_arr, full_set_res[:, 0], fh.cols)
# print '************' + fh.colsRes[1] + '***************'
# print rank_features(full_set_arr, full_set_res[:, 1], fh.cols)
kf = KFold(full_set_res.shape[0], 10)
max_sbp_corr = 0
