import numpy as np
import pandas as pd
import iisignature as isig
import seaborn as sns
sns.set_style("darkgrid")
%matplotlib inline
C:\Users\jjd\Miniconda3\lib\site-packages\IPython\html.py:14: ShimWarning: The `IPython.html` package has been deprecated since IPython 4.0. You should import from `notebook` instead. `IPython.html.widgets` has moved to `ipywidgets`. "`IPython.html.widgets` has moved to `ipywidgets`.", ShimWarning)
dim = 4
path=np.random.uniform(size=(391,dim))
signature=isig.sig(path,2)
s=isig.prepare(dim,2)
logsignature=isig.logsig(path,s)
logsignature
array([ 0.20977044, 0.31576152, -0.59549484, -0.32683967, 1.77211097, -0.98080124, -0.7372725 , -0.59238271, -1.61024298, -1.19956654])
path
array([[ 0.00599939, 0.35252989, 0.00643476], [ 0.69107634, 0.53417306, 0.13505177], [ 0.65091031, 0.68163667, 0.18822175], [ 0.40045436, 0.6050306 , 0.52894525], [ 0.69830239, 0.20724377, 0.44333588], [ 0.88958683, 0.79884177, 0.14069103], [ 0.87906374, 0.70681328, 0.94589375], [ 0.10018265, 0.04883731, 0.32006464], [ 0.31231445, 0.58446231, 0.60451132], [ 0.83650757, 0.25350303, 0.45021486], [ 0.89845868, 0.03030897, 0.98280072], [ 0.21879094, 0.62778299, 0.15134738], [ 0.39832616, 0.12370522, 0.43233098], [ 0.37466465, 0.95764397, 0.7353267 ], [ 0.04269322, 0.0144172 , 0.59703507], [ 0.82637778, 0.98277936, 0.74946826], [ 0.44112696, 0.60278035, 0.38627335], [ 0.08668496, 0.91264151, 0.39934557], [ 0.82720711, 0.09071632, 0.23124443], [ 0.30970784, 0.42012087, 0.52044559]])
dim = 2
m = 2
x = [[1,1], [3,4], [5,2], [8,6] ]
signature = isig.sig(x,m,1)
s = isig.prepare(dim,m)
logsignature = isig.logsig(x,s)
signature
(array([ 7., 5.], dtype=float32), array([ 24.5, 19. , 16. , 12.5], dtype=float32))
logsignature
array([ 7. , 5. , 1.5])
eurusd = pd.read_csv('C:/Users/jjd/Desktop/eurusd.csv', index_col=0, parse_dates=True, names=['Volume', 'Transactions'])
eurusd_by_hours = eurusd.groupby([eurusd.index.hour]).median()
eurusd_by_hours.plot(figsize=(16,6), grid=True);
audusd = pd.read_csv('C:/Users/jjd/Desktop/audusd.csv', index_col=0, parse_dates=True, names=['Volume', 'Transactions'])
audusd_by_hours = audusd.groupby([audusd.index.hour]).median()
audusd_by_hours.plot(figsize=(16,6), grid=True);
usdjpy = pd.read_csv('C:/Users/jjd/Desktop/usdjpy.csv', index_col=0, parse_dates=True, names=['Volume', 'Transactions'])
usdjpy_by_hours = usdjpy.groupby([usdjpy.index.hour]).median()
usdjpy_by_hours.plot(figsize=(16,6), grid=True);
gbpusd = pd.read_csv('C:/Users/jjd/Desktop/gbpusd.csv', index_col=0, parse_dates=True, names=['Volume', 'Transactions'])
gbpusd_by_hours = gbpusd.groupby([gbpusd.index.hour]).median()
gbpusd_by_hours.plot(figsize=(16,6), grid=True);
usdchf = pd.read_csv('C:/Users/jjd/Desktop/usdchf.csv', index_col=0, parse_dates=True, names=['Volume', 'Transactions'])
usdchf_by_hours = usdchf.groupby([usdchf.index.hour]).median()
usdchf_by_hours.plot(figsize=(16,6), grid=True);
eurjpy = pd.read_csv('C:/Users/jjd/Desktop/eurjpy.csv', index_col=0, parse_dates=True, names=['Volume', 'Transactions'])
eurjpy_by_hours = eurjpy.groupby([eurjpy.index.hour]).median()
eurjpy_by_hours.plot(figsize=(16,6), grid=True);
audjpy = pd.read_csv('C:/Users/jjd/Desktop/audjpy.csv', index_col=0, parse_dates=True, names=['Volume', 'Transactions'])
audjpy_by_hours = audjpy.groupby([audjpy.index.hour]).median()
audjpy_by_hours.plot(figsize=(16,6), grid=True);