#joint probability for hpi and mortgage in a time window (for seasonal data?)
## calculate the string of hpi and mortgage
hpi_seas_str, _ = gt.connotation([hpi_seasonal], connotation)
hpi_non_seas_str, _ = gt.connotation([hpi_non_seas], connotation)
mortgage_str, _ = gt.connotation([unique_mortgage], connotation)
mortgage_lf_str, _ = gt.connotation([mortgage_lf], connotation)
##find matches of up in hpi intervals
cfg1 = {
"pre_processing": "~ 10",
"connotation": "D1 0.005",
"expression": "p+"
}
matches_hpi = gt.ssts(hpi_seasonal, cfg1, 'clean')
## determinate the frequency of each character per window
data_freq_test = gt.CharFreq(hpi_seas_str, 1, 10)
data_freq_test2 = gt.CharFreq(mortgage_str, 1, 10)
p_joint_prob = np.multiply(data_freq_test["p"], data_freq_test2["p"])
ax1 = plt.subplot(2, 1, 1)
ax2 = plt.subplot(2, 1, 2)
# ax3 = plt.subplot(3,1,3)
ax1.plot(time_all, hpi_values)
[
ax1.axvspan(time_all[i[0]],
time_all[i[1] - 1],
#SSTS------------------------
#configure for hpi series
cfg1 = {
"pre_processing": "RM 25",
"connotation": "D1 0.05",
"expression": "p+"
}
#configure for mortgage series (only to have derivative string)
cfg2 = {"pre_processing": "", "connotation": "D1 0.01", "expression": ""}
print(np.where(time_hpi == time_all[0])[0])
hpi_values = list(hpi_values)[np.where(
time_hpi == time_all[0])[0][0]:np.where(time_hpi == time_all[-1])[0][0] +
1]
matches = gt.ssts(hpi_values, cfg1)
matches2 = gt.ssts(unique_mortgage, cfg2, report="full")
fig, (ax1, ax2, ax3) = plt.subplots(nrows=3, ncols=1)
ax1.plot(time_all, hpi_values)
[
ax1.axvspan(time_all[i[0]], time_all[i[1]], facecolor='green', alpha=0.10)
for i in matches
]
ax1.set_title("HPI")
ax1.set_xlabel("date (by month)")
ax2.plot(time_all, unique_mortgage)
max_p = []
sorted(tuple_list, key=lambda i: i[1]), lambda i: i[1])]
m, t = ([i[0] for i in tuple_list], [i[1] for i in tuple_list])
time = sorted(list(set(time_hpi_1).intersection(t)))
hpi_values = hpi_values1[[time_hpi_1.index(x) for x in time]]
cfg = {"pre_processing": "", "connotation": "D1 0.01", "expression": "p+"}
cfg2 = {
"pre_processing": "HP 10",
"connotation": "D1 0.01",
"expression": "p+"
}
matches = ssts.ssts(hpi_values, cfg2)
matches2 = ssts.ssts(hpi_values1, cfg, report="full")
fig, axs = plt.subplots(nrows=2, ncols=1, sharex=True)
fig.suptitle("HPI and Mortgage")
axs[0].plot(time, hpi_values)
axs[1].set_xlabel("Date")
axs[0].set_ylabel("House Price Index")
[
axs[0].axvspan(time[i[0]], time[i[1] - 1], facecolor='green', alpha=0.10)
for i in matches
]
axs[1].plot(t, m)
axs[1].set_ylabel("Mortgage Counts")
"expression": "[10].(1z){1000,}"
}
i = 0
for tag, iotip, cfg, lag in zip(["wrist", "hand", "elbow", "torso"],
[iotip_wrist, iotip_hand, iotip_elbow, torso],
[cfg1, cfg1, cfg2, cfg1], [1, 0, 0, 0]):
acc_[tag] = load_sensor_IOTIP(iotipsPath + iotip, "Green", "Accelerometer")
gyr_[tag] = load_sensor_IOTIP(iotipsPath + iotip, "Green", "Gyroscope")
mag_[tag] = load_sensor_IOTIP(iotipsPath + iotip, "Green", "Magnetometer")
#Align Data
time[tag], acc_[tag], gyr_[tag], mag_[tag] = syncData(
acc_[tag], gyr_[tag], mag_[tag])
matches = gt.ssts(acc_[tag][:10000, 0], cfg)
delay = matches[lag][0]
time[tag] = time[tag][delay:]
acc_[tag] = acc_[tag][delay:]
gyr_[tag] = gyr_[tag][delay:]
mag_[tag] = mag_[tag][delay:]
#sync devices
time, acc_, gyr_, mag_ = syncDevicesData(acc_, gyr_, mag_)
#Pre_process
# acc = lowpass(acc_, f=25, fs=100)
# gyr = lowpass(gyr_, f=25, fs=100)
# mag = lowpass(mag_, f=25, fs=100)
}
cfg3 = {
"pre_processing": "",
"connotation": "D1 0.1",
"expression": "n[zp]"
}
cfg4 = {"pre_processing": "", "connotation": "D1 0.05", "expression": "z+"}
print(final)
#
# matches1 = gt.ssts(final, cfg1)
# matches2 = gt.ssts(final, cfg2)
# matches3 = gt.ssts(final, cfg3)
matches4 = gt.ssts(final, cfg4)
plt.plot(sig)
plt.plot(sig_std + 2)
plt.plot(sig_sum + 4)
plt.plot(sig_zcr + 6)
plt.plot(sig_ad + 8)
plt.plot(final + 10)
# [plt.axvline(i[0] + (i[1]-i[0])//2) for i in matches1]
# [plt.axvline(i[0] + (i[1]-i[0])//2) for i in matches2]
# [plt.axvline(i[0] + (i[1]-i[0])//2) for i in matches3]
[plt.axvline(i[0] + (i[1] - i[0]) // 2) for i in matches4]
#[plt.axvspan(i[0]-2, i[1]+2, facecolor='green', alpha=1) for i in matches1]