def plot_acf_pacf(timeSeries,
nlags=40,
alpha=0.05,
method_acf=True,
method_pacf='ywunbiased',
legend=["ACF", "PACF"],
labels=["timeSeries"]):
valuesACF, confIntACF = tsa.acf(timeSeries,
nlags=30,
alpha=0.05,
unbiased=True,
qstat=False)
# qstat: For LJung-Box values and pvalues: For Pvalues of this
## For some reason the first value of the PACF is 1, when it should not be defined
valuesPACF, confIntPACF = tsa.pacf(timeSeries,
nlags=30,
alpha=0.05,
method='ywunbiased')
gl.set_subplots(2, 1)
gl.stem([],
valuesACF,
labels=[labels[0], "lag", "ACF"],
legend=[legend[0]])
# gl.plot([],confIntACF, nf = 0, color = "blue",
# legend = ["ConfInt %.2f" % (1 - alpha)],
# lw = 1)
plt.axhline(y=-1.96 / np.sqrt(len(timeSeries)),
linestyle='--',
color='gray',
lw=3)
plt.axhline(y=1.96 / np.sqrt(len(timeSeries)),
linestyle='--',
color='gray',
lw=3)
gl.stem([],
valuesPACF,
labels=[labels[0], "lag", "PACF"],
legend=[legend[1]])
# gl.plot([],confIntPACF, nf = 0, color = "blue",
# legend = ["ConfInt %.2f" % (1 - alpha)])
plt.axhline(y=-1.96 / np.sqrt(len(timeSeries)),
linestyle='--',
color='gray',
lw=3)
plt.axhline(y=1.96 / np.sqrt(len(timeSeries)),
linestyle='--',
color='gray',
lw=3)
def plot_acf_pacf(timeSeries, nlags = 40, alpha = 0.05,
method_acf = True, method_pacf = 'ywunbiased',
legend = ["ACF", "PACF"],
labels = ["timeSeries"]):
valuesACF, confIntACF = tsa.acf(timeSeries, nlags = 30,
alpha = 0.05, unbiased = True,
qstat = False)
# qstat: For LJung-Box values and pvalues: For Pvalues of this
## For some reason the first value of the PACF is 1, when it should not be defined
valuesPACF, confIntPACF = tsa.pacf(timeSeries, nlags = 30,
alpha = 0.05,
method = 'ywunbiased')
gl.set_subplots(2,1)
gl.stem([],valuesACF,
labels = [labels[0], "lag", "ACF"],
legend = [legend[0]])
# gl.plot([],confIntACF, nf = 0, color = "blue",
# legend = ["ConfInt %.2f" % (1 - alpha)],
# lw = 1)
plt.axhline(y=-1.96/np.sqrt(len(timeSeries)),
linestyle='--',color='gray', lw = 3)
plt.axhline(y=1.96/np.sqrt(len(timeSeries)),
linestyle='--',color='gray', lw = 3)
gl.stem([],valuesPACF,
labels = [labels[0], "lag", "PACF"],
legend = [legend[1]])
# gl.plot([],confIntPACF, nf = 0, color = "blue",
# legend = ["ConfInt %.2f" % (1 - alpha)])
plt.axhline(y=-1.96/np.sqrt(len(timeSeries)),
linestyle='--',color='gray', lw = 3)
plt.axhline(y=1.96/np.sqrt(len(timeSeries)),
linestyle='--',color='gray', lw = 3)
def get_TMA(time_series, L):
""" First it trains the data so that the prediction is maximized"""
### Training phase, we obtained the MSQE of the filter for predicting the next value
time_series = time_series.flatten()
Ns = time_series.size
Xtrain, Ytrain = ul.windowSample(time_series, L)
window = np.linalg.pinv((Xtrain.T).dot(Xtrain))
window = window.dot(Xtrain.T).dot(Ytrain)
window = np.fliplr([window])[0]
gl.stem([], window)
sM = np.convolve(time_series.flatten(),window.flatten(), mode = "full")
# print sM.shape
sM = sM
sM = sM/np.sum(window) # Divide so that it is the actual mean.
sM[:L] = (np.ones((L,1)) * sM[L]).flatten() # Set the first ones equal to the first fully obtained stimator
sM = sM[:-L+1] # Remove the last values since they hare convolved with 0's as well
return sM
dates = Cartera.get_timeData(symbolIDs[symbol_ID_indx1],periods[0]).get_dates()
#print dates[0], dates[26], dates[27]
################# Plotting the data #################
if(trading_graph):
# Trading plot of the points !
title = "CLOSE Price, Volume and Return evolution"
ax1 = gl.subplot2grid((5,1), (0,0), rowspan=2, colspan=1)
gl.tradingBarChart(Cartera.get_timeData(symbolIDs[symbol_ID_indx1],periods[0]), ax = ax1,
dataTransform = dataTransform, AxesStyle = "Normal - No xaxis",
labels = [title,"",symbolIDs[symbol_ID_indx1] +"(" +str(periods[0])+")"])
ax2 = gl.subplot2grid((5,1), (2,0), rowspan=1, colspan=1, sharex = ax1)
gl.stem(dates, volume, ax = ax2, dataTransform = dataTransform,
AxesStyle = "Normal - No xaxis - Ny:4",
labels = ["","",symbolIDs[0] +"("+ str(periods[0])+ "M)"], legend = [ "Volume"])
ax3 = gl.subplot2grid((5,1), (3,0), rowspan=2, colspan=1, sharex = ax1)
gl.stem(dates, ret1, ax = ax3, dataTransform = dataTransform,
AxesStyle = "Normal",
labels = ["","",symbolIDs[0] +"("+ str(periods[0])+ "M)"], legend = ["Return"])
#
gl.set_fontSizes(ax = [ax1,ax2,ax3], title = 20, xlabel = 20, ylabel = 20,
legend = 20, xticks = 10, yticks = 10)
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.01, hspace=0.01)
gl.savefig(folder_images +'PriceAndReturns1Symbol_EM.png',
dpi = 100, sizeInches = [22, 12])
alpha_stem = 0.5
marker_stem = [".",1,None]
gl.init_figure();
ax1 = gl.subplot2grid((4,1), (0,0), rowspan=1, colspan=1)
ax2 = gl.subplot2grid((4,1), (1,0), rowspan=1, colspan=1, sharex = ax1)
ax3 = gl.subplot2grid((4,1), (2,0), rowspan=1, colspan=1, sharex = ax1)
ax4 = gl.subplot2grid((4,1), (3,0), rowspan=1, colspan=1, sharex = ax1)
## Ax1 = Close price at the end of the sessions
gl.plot(days_keys, C, ax = ax1, labels = ["Time variables","","Close Rate"],
AxesStyle = "Normal - No xaxis", legend = ["Close Price"], color = "k")
## Ax2 = days
gl.stem(days_keys, day_of_week, ax = ax2, labels = ["","","Day"], bottom = 0.0,
AxesStyle = "Normal - No xaxis", alpha = alpha_stem,
marker = marker_stem, color = "k", legend = ["Day"])
# Ax3 = week
gl.stem(days_keys,week_of_year ,bottom = 0.0, ax = ax3, labels = ["","","Week"],AxesStyle = "Normal - No xaxis",
alpha =alpha_stem, marker = marker_stem, color = "k", legend = ["Week"])
## Ax4 = year
gl.plot(days_keys,year, ax = ax4, labels = ["","","Year"],AxesStyle = "Normal",
alpha = alpha_stem, color = "k", legend = ["Year"])
# Set final properties and save figure
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.05, hspace=0.05)
gl.set_fontSizes(ax = [ax1,ax2,ax3,ax4], title = 20, xlabel = 20, ylabel = 20,
legend = 15, xticks = 12, yticks = 12)
############## Average and Window ################
ax1 = gl.subplot2grid((1, 5), (0, 0), rowspan=1, colspan=3)
title = "Price and SMA. " + str(
symbols[0]) + "(" + ul5.period_dic[timeData.period] + ")"
gl.plot(dates, [price, SMA],
AxesStyle="Normal",
labels=[title, "", r"Price ($\$$)"],
legend=["Price", "SMA(%i)" % nMA1])
ax2 = gl.subplot2grid((1, 5), (0, 3), rowspan=1, colspan=2)
gl.stem([],
SMAw,
nf=0,
AxesStyle="Normal2",
labels=["SMA Window", "lag", ""],
legend=["SMAw(%i)" % nMA1],
xlimPad=[0.1, 0.3],
ylimPad=[0.1, 0.4],
marker=[".", 10, None])
gl.subplots_adjust(left=.09,
bottom=.10,
right=.90,
top=.95,
wspace=.35,
hspace=0)
gl.savefig(folder_images + 'basicSMA.png', dpi=100, sizeInches=[14, 5])
""" 2nd GRAPH """
############## PLOTTING ################
gl.set_subplots(3, 2)
Y = (np.sign(np.diff(X,n=1,axis = 0)) +1)/2.0
Y_noisy = (np.sign(np.diff(X_noisy,n=1,axis = 0)) +1)/2.0
## Subselect both the signals X and the labels Y since we cannot know the prediction of the last sample.
if (plot_labelling_signal and plot_flag):
gl.set_subplots(2,1)
# Plot the original signal and its labelling !
ax1 = gl.scatter(tgrid,X, lw = 1, alpha = 0.9, color = "k", nf = 1,
labels = ["Original signal","",r"$\mu(t)$"])
gl.plot(tgrid,X, lw = 2, color = "k", ls = "--")
gl.set_fontSizes( title = 20, xlabel = 15, ylabel = 20,
legend = 12, xticks = 12, yticks = 12)
gl.stem(tgrid[:-1,0], Y, sharex = ax1, nf = 1, labels = ["", "t","Y(t)"], bottom = 0.5)
# Plot noisy part
# gl.scatter(tgrid,X_noisy, lw = 1, alpha = 0.9, color = "k", nf = 1,
# labels = ["Noisy signal","","X(t)"])
# gl.plot(tgrid,X_noisy, lw = 2, color = "k", ls = "--")
# gl.stem(tgrid[:-1,0], Y_noisy, sharex = ax1, nf = 1, labels = ["", "t","Y_noise(t)"])
#
gl.set_fontSizes( title = 20, xlabel = 15, ylabel = 20,
legend = 12, xticks = 12, yticks = 12)
gl.savefig(folder_images +'GP_labelling_of_the_signal.png',
dpi = 100, sizeInches = [2*6.5, 2*2])
## It would be nice to see !!
# Generate a lot of random signals and compute the mean prediction of Y
SMA = timeData.SMA(n = nMA1)
""" 1st GRAPH """
############## Average and Window ################
ax1 = gl.subplot2grid((1,5), (0,0), rowspan=1, colspan=3)
title = "Price and SMA. " + str(symbols[0]) + "(" + ul.period_dic[timeData.period]+ ")"
gl.plot(dates, [price, SMA] ,AxesStyle = "Normal",
labels = [title,"",r"Price ($\$$)"],
legend = ["Price", "SMA(%i)"%nMA1])
ax2 = gl.subplot2grid((1,5), (0,3), rowspan=1, colspan=2)
gl.stem([], SMAw, nf = 0, AxesStyle = "Normal2",
labels = ["SMA Window","lag",""],
legend = ["SMAw(%i)"%nMA1],
xlimPad = [0.1,0.3], ylimPad = [0.1,0.4],
marker = [".",10,None])
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.35, hspace=0)
gl.savefig(folder_images +'basicSMA.png',
dpi = 100, sizeInches = [14,5])
""" 2nd GRAPH """
############## PLOTTING ################
gl.set_subplots(3,2)
marker = [".",10,None]
# Plotting the 3 of them at the same time.
ax1 = gl.stem([], SMAw, nf = 1,
labels = ["Windows L = %i"%nMA1,"","SMA"],
legend = ["SMA(%i)"%nMA1],
gl.barchart(dates, dataHLOC)
gl.plot(dates, EMAslow, legend=["Slow"], lw=1, color="b")
gl.plot(dates, EMAfast, legend=["fast"], lw=1, color="r", xaxis_mode="hidden")
## Entry Axes
ax2 = gl.plot(dates,
ul.scale(EMAfast - EMAslow),
legend=["Difference"],
nf=1,
sharex=ax1,
labels=["", "", "Events"],
fill=1,
alpha=0.3)
gl.stem(dates, crosses, legend=["TradeSignal"])
gl.scatter(dates, ul.scale(EMAfast - EMAslow), lw=0.5, alpha=0.5)
gl.plot(dates, np.zeros(crosses.shape))
## Exit Axes
gl.add_hlines(datesExit,
all_stops,
ax=ax1,
legend=["TrailingStop"],
alpha=0.8,
lw=0.2)
ax3 = gl.plot(datesExit,
ul.scale(-(all_stops - pCheckCross)),
legend=["TrailingStop"],
nf=1,
sharex=ax1,
title = "CLOSE Price, Volume and Return evolution"
ax1 = gl.subplot2grid((4,1), (0,0), rowspan=1, colspan=1)
gl.tradingBarChart(Cartera.get_timeData(symbolIDs[symbol_ID_indx1],periods[0]), ax = ax1,
dataTransform = dataTransform, AxesStyle = "Normal - No xaxis", color = "k",
labels = [title,"",symbolIDs[symbol_ID_indx1] +"(" +str(periods[0])+"M)"])
ax2 = gl.subplot2grid((4,1), (1,0), rowspan=1, colspan=1, sharex = ax1)
gl.tradingBarChart(Cartera.get_timeData(symbolIDs[symbol_ID_indx2],periods[0]), ax = ax2,
dataTransform = dataTransform, AxesStyle = "Normal - No xaxis", color = "k",
labels = ["","",symbolIDs[symbol_ID_indx2] +"(" +str(periods[0])+"M)"])
ax3 = gl.subplot2grid((4,1), (2,0), rowspan=1, colspan=1)
gl.stem(dates, ret1, ax = ax3, dataTransform = dataTransform,
AxesStyle = "Normal",
labels = ["","",symbolIDs[symbol_ID_indx1] +"("+ str(periods[0])+ "M)"], legend = ["Return"])
ax4 = gl.subplot2grid((4,1), (3,0), rowspan=1, colspan=1, sharex = ax1, sharey = ax3)
gl.stem(dates, ret2, ax = ax4, dataTransform = dataTransform,
AxesStyle = "Normal",
labels = ["","",symbolIDs[symbol_ID_indx2] +"("+ str(periods[0])+ "M)"], legend = ["Return"])
#
gl.set_fontSizes(ax = [ax1,ax2,ax3,ax4], title = 20, xlabel = 20, ylabel = 20,
legend = 20, xticks = 10, yticks = 10)
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.01, hspace=0.01)
gl.savefig(folder_images +'PriceAndReturns2Symbol_EM.png',
dpi = 100, sizeInches = [22, 12])
type_graph = 2
folder_images = "../pics/gl/"
dpi = 100
sizeInches = [2*8, 2*3]
# Subplot Type 1
if (type_graph == 1):
gl.set_subplots(nr = 1, nc = 2)
gl.plot(X,Y, nf = 1,
color = "k", lw = 5, alpha = 0.7,
labels = ["Sine chart","Time (s)", "Voltage(V)"],
legend = ["Rolling measurement"])
gl.stem(X2,Y2, nf = 1,
color = "k", lw = 2, alpha = 0.7,
labels = ["Discrete window","Sample (k)", "Amplitud"],
legend = ["Window values"])
gl.savefig(folder_images +'subplot1.png',
dpi = dpi, sizeInches = sizeInches)
# Subplot Type 2
if (type_graph == 2):
ax1 = gl.subplot2grid((1,4), (0,0), rowspan=1, colspan=3)
gl.plot(X,Y, nf = 0,
color = "k", lw = 5, alpha = 0.7,
labels = ["Sine chart","Time (s)", "Voltage(V)"],
legend = ["Rolling measurement"])
#
ax2 = gl.subplot2grid((1,4), (0,3), rowspan=1, colspan=1)
gl.plot(X2,Y2, nf = 0,
color = "k", lw = 2, alpha = 0.7,
########## Plotting ! ####################
gl.set_subplots(3,1)
## Price Axes
ax1 = gl.plot(dates,price, legend = ["Price"],
labels = ["Crossing MA Strategy", "","Price"], alpha = 0, nf = 1)
gl.barchart(dates, dataHLOC)
gl.plot(dates,EMAslow, legend = ["Slow"], lw = 1, color = "b")
gl.plot(dates,EMAfast, legend = ["fast"], lw = 1, color = "r", xaxis_mode = "hidden")
## Entry Axes
ax2 = gl.plot(dates, ul.scale(EMAfast - EMAslow), legend = ["Difference"]
,nf = 1, sharex = ax1, labels = ["","","Events"], fill = 1, alpha = 0.3)
gl.stem(dates,crosses, legend = ["TradeSignal"] )
gl.scatter(dates, ul.scale(EMAfast - EMAslow), lw = 0.5, alpha = 0.5)
gl.plot(dates,np.zeros(crosses.shape))
## Exit Axes
gl.add_hlines(datesExit, all_stops, ax = ax1, legend = ["TrailingStop"], alpha = 0.8, lw = 0.2)
ax3 = gl.plot(datesExit, ul.scale(-(all_stops - pCheckCross)), legend = ["TrailingStop"]
,nf = 1, sharex = ax1, labels = ["","","TrailingStop"], lw = 0.5, alpha = 0.5, fill = 1)
gl.stem(datesExit,exitcrosses, legend = ["TradeSignal"] )
gl.scatter(datesExit, ul.scale(-(all_stops - pCheckCross)), lw = 0.5, alpha = 0.5)
gl.add_hlines(datesExit2, all_stops2, ax = ax1, legend = ["TrailingStop2"], alpha = 0.8, lw = 0.2)
gl.plot(datesExit2, ul.scale(all_stops2 - pCheckCross2), legend = ["TrailingStop2"],
lw = 0.5, alpha = 0.5, fill = 1)
gl.stem(datesExit2,exitcrosses2, legend = ["TradeSignal"] )
gl.scatter(datesExit2, ul.scale(all_stops2 - pCheckCross2), lw = 0.5, alpha = 0.5)
marker_stem = [".",1,None]
train_color = "b"
test_color = "r"
gl.init_figure();
ax1 = gl.subplot2grid((4,1), (0,0), rowspan=1, colspan=1)
ax2 = gl.subplot2grid((4,1), (1,0), rowspan=1, colspan=1, sharex = ax1)
ax3 = gl.subplot2grid((4,1), (2,0), rowspan=1, colspan=1, sharex = ax1)
ax4 = gl.subplot2grid((4,1), (3,0), rowspan=1, colspan=1, sharex = ax1)
## Ax1 = Close price at the end of the sessions
gl.plot(days_keys, C, ax = ax1, labels = ["Results " + key_classifier,"","Close Rate"],
AxesStyle = "Normal - No xaxis", legend = ["Close Rate"])
## Ax2 = 1 if the stock has gone up, zero if it has gone down
gl.stem(dates_train, Ytrain_reg, ax = ax2, labels = ["","","Target_reg"], bottom = 0.0, AxesStyle = "Normal - No xaxis", alpha = alpha_stem,
marker = marker_stem, color = train_color, legend = ["tr"])
gl.stem(dates_test, Ytest_reg, ax = ax2, labels = ["","","Target_reg"],bottom = 0.0,AxesStyle = "Normal - No xaxis", color = test_color ,
alpha = alpha_stem, marker = marker_stem, legend = ["tst"])
# Ax3 = The estimates probability for train and test
gl.stem(dates_train,Ypredict_train_proba ,bottom = 0.5, ax = ax3, labels = ["","",r"$\hat{Y}$"],AxesStyle = "Normal - No xaxis",
alpha =alpha_stem, marker = marker_stem, color = train_color, legend = ["tr"])
gl.stem(dates_test, Ypredict_test_proba,bottom = 0.5, ax = ax3, labels = ["","",r"$\hat{Y}$"],color = test_color,ylim = [0.0,1.0],alpha = alpha_stem,
AxesStyle = "Normal - No xaxis", marker = marker_stem, legend = ["tst"]) # ylim = [0.0,1.0]
## Ax4 = Cross Entropy of the samples for train and test
gl.stem(dates_train,train_cross_entropy ,bottom = 0.0, ax = ax4, labels = ["","","CE"],AxesStyle = "Normal",
alpha = alpha_stem, marker = marker_stem, color = train_color, legend = ["tr"])
gl.stem(dates_test, test_cross_entropy,bottom = 0.0, ax = ax4, color = test_color,AxesStyle = "Normal",
alpha = alpha_stem,marker = marker_stem, legend = ["tst"])
Returns = timeData.get_timeSeriesReturn(seriesNames = ["Close"]);
CumReturns = timeData.get_timeSeriesCumReturn(seriesNames = ["Close"]);
nSMA = 10
SMA = timeData.SMA(n = nSMA)
ax1 = gl.plot(dates, SMA, labels = [timeData.symbolID + str(timeData.period), "Time", "Price"],
legend = ["SMA(%i)" % nSMA], nf = 1, dataTransform = dataTransform,
AxesStyle = "Normal - No xaxis", color = "cobalt blue")
gl.barchart(dates, dataHLOC, lw = 2, dataTransform = dataTransform, color = "k",
AxesStyle = "Normal - No xaxis", legend = ["OHCL"])
# Notice the Weekends and the displacement between bars and step
############# 2: Volume
gl.stem(dates, volume, sharex = ax1, labels = ["", "Time", "Volume"],
legend = ["Volume"], nf = 1, alpha = 0.8, dataTransform = dataTransform,
AxesStyle = "Normal - No xaxis")
############# 3: Returns
gl.stem(dates, Returns, sharex = ax1, labels = ["", "Time", "Return"],
legend = ["Return"], nf = 1, dataTransform = dataTransform,
AxesStyle = "Normal - No xaxis")
############# 4: Commulative Returns
seriesNames = ["Close"]
gl.plot(dates, CumReturns, sharex = ax1,labels = ["", "Time", "Cum Return"],
legend = ["Cum Return"], nf = 1, dataTransform = dataTransform,
AxesStyle = "Normal")
############## Dayly things obtaining ####################################
# We separate the data into a list of days in order to be able to analyze it easier
ax1 = gl.subplot2grid((5, 1), (0, 0), rowspan=2, colspan=1)
gl.tradingBarChart(
Cartera.get_timeData(symbolIDs[symbol_ID_indx1], periods[0]),
ax=ax1,
dataTransform=dataTransform,
AxesStyle="Normal - No xaxis",
labels=[
title, "", symbolIDs[symbol_ID_indx1] + "(" + str(periods[0]) + ")"
])
ax2 = gl.subplot2grid((5, 1), (2, 0), rowspan=1, colspan=1, sharex=ax1)
gl.stem(dates,
volume,
ax=ax2,
dataTransform=dataTransform,
AxesStyle="Normal - No xaxis - Ny:4",
labels=["", "", symbolIDs[0] + "(" + str(periods[0]) + "M)"],
legend=["Volume"])
ax3 = gl.subplot2grid((5, 1), (3, 0), rowspan=2, colspan=1, sharex=ax1)
gl.stem(dates,
ret1,
ax=ax3,
dataTransform=dataTransform,
AxesStyle="Normal",
labels=["", "", symbolIDs[0] + "(" + str(periods[0]) + "M)"],
legend=["Return"])
#
gl.set_fontSizes(ax=[ax1, ax2, ax3],
title=20,
############################################################
periods = mySymbol.get_periods()
gl.set_subplots(4, 1, sharex=True)
# TODO: Be able to automatize the shareX thing
for period in periods:
myTimeData = mySymbol.get_timeDataObject(period)
price = myTimeData.get_timeSeries(["Close"])
volume = myTimeData.get_timeSeries(["Volume"])
dates = myTimeData.get_dates()
gl.plot(dates,
price,
labels=["", "", mySymbol.symbol],
legend=[str(period)])
gl.stem(dates, volume, legend=[str(period)], na=1, nf=0)
gl.subplots_adjust(left=.09,
bottom=.10,
right=.90,
top=.95,
wspace=.20,
hspace=0)
# TODO: We can also use the library of indicators especifing the period.
# If no period specified, we use the dayly or the bigges one.
#######################
## Ax1 = Close price at the end of the sessions
gl.plot(days_keys,
C,
ax=ax1,
labels=["Time variables", "", "Close Rate"],
AxesStyle="Normal - No xaxis",
legend=["Close Price"],
color="k")
## Ax2 = days
gl.stem(days_keys,
day_of_week,
ax=ax2,
labels=["", "", "Day"],
bottom=0.0,
AxesStyle="Normal - No xaxis",
alpha=alpha_stem,
marker=marker_stem,
color="k",
legend=["Day"])
# Ax3 = week
gl.stem(days_keys,
week_of_year,
bottom=0.0,
ax=ax3,
labels=["", "", "Week"],
AxesStyle="Normal - No xaxis",
alpha=alpha_stem,
marker=marker_stem,
color="k",
dates = timeData.get_dates()
df = timeData.get_timeData()
# Momentum and Rate of convergence
nAD = 5
ACCDIST = timeData.ACCDIST(n=nAD)
# Plotting
gl.set_subplots(2, 1)
gl.plot(dates,
price,
nf=1,
labels=["Momentum Indicators ACC DIST", "", "Price"],
legend=["Price", " Momentum", "ROC"])
vol = timeData.get_timeSeries(["Volume"])
gl.stem(dates, vol, nf=0, na=1, legend=["Volume"])
gl.plot(dates, ACCDIST, nf=1, na=0, legend=["ACCDIST(%i)" % nMOM])
gl.plot(dates, np.zeros((dates.size, 1)), nf=0, na=0)
gl.subplots_adjust(left=.09,
bottom=.10,
right=.90,
top=.95,
wspace=.20,
hspace=0)
gl.savefig(folder_images + 'AC.png', dpi=100, sizeInches=[2 * 8, 2 * 2])
#########################################################
# Oscillators
# Set the timeSeries to operate with.
mySymbol.set_seriesNames(["Close"])
############################################################
###### NOW WE OPERATE DIRECTLY ON THE TIMEDATAs ############
############################################################
periods = mySymbol.get_periods()
gl.set_subplots(4,1, sharex = True)
# TODO: Be able to automatize the shareX thing
for period in periods:
myTimeData = mySymbol.get_timeDataObject(period)
price = myTimeData.get_timeSeries(["Close"])
volume = myTimeData.get_timeSeries(["Volume"])
dates = myTimeData.get_dates()
gl.plot(dates, price, labels = ["", "", mySymbol.symbol],
legend = [str(period)])
gl.stem(dates, volume,
legend = [str(period)], na = 1, nf = 0)
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.20, hspace=0)
# TODO: We can also use the library of indicators especifing the period.
# If no period specified, we use the dayly or the bigges one.
#######################
EMAw = indl.get_EMA(delta, nHMA, cval = 1)
dEMAw = bMA.convolve(EMAw, diff_final)
# EMAEMA = bMA.diff(EMAEMA, cval = np.nan)
# EMAEMA = indl.get_EMA(diffw1, nHMA, cval = 1)
## Plotting just the 3 windows
gl.set_subplots(1,3)
marker = [".",20,"r"]
lw = 4
# Plotting the 3 of them at the same time.
ax1 = gl.stem([], diffw1, nf = 1, lw = lw,
labels = ["MOMw(%i)"%ndiff1,"lag","Value"],
xlimPad = [0.1,0.3], ylimPad = [0.1,0.1],
marker = marker, AxesStyle = "Normal2")
gl.stem([], diffw2, nf = 1, sharex = ax1, sharey = ax1,lw = lw,
labels = ["MOMw(%i)"%ndiff2,"lag",""],
xlimPad = [0.1,0.3], ylimPad = [0.1,0.1],
marker = marker, AxesStyle = "Normal2 - No yaxis")
gl.stem([], diffw3, nf = 1,sharex = ax1, sharey = ax1,lw = lw,
labels = ["MOMw(%i)"%ndiff3,"lag",""],
xlimPad = [0.1,0.3], ylimPad = [0.1,0.1],
marker = marker, AxesStyle = "Normal2 - No yaxis")
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.050, hspace=0.01)
gl.savefig(folder_images +'MOMw.png',
alpha=0.9,
color="k",
nf=1,
labels=["Original signal", "", r"$\mu(t)$"])
gl.plot(tgrid, X, lw=2, color="k", ls="--")
gl.set_fontSizes(title=20,
xlabel=15,
ylabel=20,
legend=12,
xticks=12,
yticks=12)
gl.stem(tgrid[:-1, 0],
Y,
sharex=ax1,
nf=1,
labels=["", "t", "Y(t)"],
bottom=0.5)
# Plot noisy part
# gl.scatter(tgrid,X_noisy, lw = 1, alpha = 0.9, color = "k", nf = 1,
# labels = ["Noisy signal","","X(t)"])
# gl.plot(tgrid,X_noisy, lw = 2, color = "k", ls = "--")
# gl.stem(tgrid[:-1,0], Y_noisy, sharex = ax1, nf = 1, labels = ["", "t","Y_noise(t)"])
#
gl.set_fontSizes(title=20,
xlabel=15,
ylabel=20,
legend=12,
xticks=12,
yticks=12)
x = np.array(range(C.size)) + cumulated_samples
cumulated_samples += C.size
gl.plot(x,
C,
ax=ax1,
legend=legend,
color="k",
alpha=alpha,
labels=[title, "", r"Rate"],
AxesStyle="Normal - No xaxis")
gl.stem(x,
V,
ax=ax2,
legend=legend,
color="k",
alpha=alpha,
labels=[title, "", r"Rate"],
AxesStyle="Normal - No xaxis")
gl.set_fontSizes(ax=[ax1],
title=20,
xlabel=20,
ylabel=20,
legend=10,
xticks=10,
yticks=10)
image_name = "dayly_examples"
gl.savefig(folder_images + image_name, dpi=100, sizeInches=[20, 7])
print ("Computing batch %i/%i"%(step, model.input.epoch_size))
vals = session.run(fetches, feed_dict)
state = vals["final_state"]
output = vals["output"]
input_i = vals["input"]
outputs.append(output)
inputs.append(input_i)
targets.append(vals["targets"])
if (step == 100):
break;
if (plot_data):
batch_i = 5
data = np.array(inputs[batch_i][0])[:,[0]]
labels = np.array(targets[batch_i][0])[:]
predicted = np.array(outputs[batch_i][0])[:,[1]]
#print(data)
#print(labels)
#print (predicted)
labels_chart = ["Example output for medium noise level", "","X[n]"]
gl.set_subplots(3,1)
ax1 = gl.plot(np.array(range(data.size)), data, nf = 1, labels = labels_chart, legend = ["X[n]"])
ax2 = gl.stem(np.array(range(data.size)),labels, nf = 1, sharex = ax1, labels = ["","","Y[n]"], bottom = 0.5,
legend = ["Targets Y[n]"])
gl.stem(np.array(range(data.size)),predicted, nf = 1, sharex = ax1, sharey = ax2, labels = ["","n","O[n]"],bottom = 0.5,
legend = ["Predictions O[n]"])
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.2, hspace=0.01)
title = "Price evolution for 2 securities"
ax1 = gl.tradingBarChart(Cartera.get_timeData(symbolIDs[0],15), nf = 1,
dataTransform = dataTransform, AxesStyle = "Normal - No xaxis",
labels = [title,"",symbolIDs[0] +"(15M)"])
gl.tradingBarChart(Cartera.get_timeData(symbolIDs[1],15), nf = 1, sharex = ax1,
dataTransform = dataTransform, AxesStyle = "Normal",
labels = ["","",symbolIDs[1] +"(15M)"])
gl.savefig(folder_images +'PriceEvolution2Symb15.png',
dpi = 100, sizeInches = [14, 7])
# Returns stem plot of the points
gl.set_subplots(2,1)
title = "Return for 2 securities"
ax1 = gl.stem(dates, ret1, nf = 1, dataTransform = dataTransform,
AxesStyle = "Normal - No xaxis",
labels = [title,"",symbolIDs[0] +"(15M)"])
gl.stem(dates, ret2, nf = 1, dataTransform = dataTransform, sharex = ax1,
AxesStyle = "Normal",
labels = ["","",symbolIDs[1] +"(15M)"])
gl.savefig(folder_images +'Returns2Symb15.png',
dpi = 100, sizeInches = [14, 7])
##########################################################################
################# PREPROCESS DATA ######################################
##########################################################################
## Set GAP return as NAN
gap_ret = np.where(dates.time == dates[0].time())[0]
ret1[gap_ret,:] = np.NaN
price = timeData.get_timeSeries(["Close"]);
dates = timeData.get_dates()
df = timeData.get_timeData()
# Momentum and Rate of convergence
nAD = 5
ACCDIST = timeData.ACCDIST(n = nAD)
# Plotting
gl.set_subplots(2,1)
gl.plot(dates, price , nf = 1,
labels = ["Momentum Indicators ACC DIST","","Price"],
legend = ["Price", " Momentum", "ROC"])
vol = timeData.get_timeSeries(["Volume"])
gl.stem(dates, vol , nf = 0, na = 1,
legend = ["Volume"])
gl.plot(dates, ACCDIST , nf = 1, na = 0,
legend = ["ACCDIST(%i)"%nMOM])
gl.plot(dates, np.zeros((dates.size,1)) , nf = 0, na = 0)
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.20, hspace=0)
gl.savefig(folder_images +'AC.png',
dpi = 100, sizeInches = [2*8, 2*2])
#########################################################
# Oscillators
nf=1,
sharex=ax1,
dataTransform=dataTransform,
AxesStyle="Normal",
labels=["", "", symbolIDs[1] + "(15M)"])
gl.savefig(folder_images + 'PriceEvolution2Symb15.png',
dpi=100,
sizeInches=[14, 7])
# Returns stem plot of the points
gl.set_subplots(2, 1)
title = "Return for 2 securities"
ax1 = gl.stem(dates,
ret1,
nf=1,
dataTransform=dataTransform,
AxesStyle="Normal - No xaxis",
labels=[title, "", symbolIDs[0] + "(15M)"])
gl.stem(dates,
ret2,
nf=1,
dataTransform=dataTransform,
sharex=ax1,
AxesStyle="Normal",
labels=["", "", symbolIDs[1] + "(15M)"])
gl.savefig(folder_images + 'Returns2Symb15.png',
dpi=100,
sizeInches=[14, 7])
##########################################################################
# Subplot Type 1
if (type_graph == 1):
gl.set_subplots(nr=1, nc=2)
gl.plot(X,
Y,
nf=1,
color="k",
lw=5,
alpha=0.7,
labels=["Sine chart", "Time (s)", "Voltage(V)"],
legend=["Rolling measurement"])
gl.stem(X2,
Y2,
nf=1,
color="k",
lw=2,
alpha=0.7,
labels=["Discrete window", "Sample (k)", "Amplitud"],
legend=["Window values"])
gl.savefig(folder_images + 'subplot1.png', dpi=dpi, sizeInches=sizeInches)
# Subplot Type 2
if (type_graph == 2):
ax1 = gl.subplot2grid((1, 4), (0, 0), rowspan=1, colspan=3)
gl.plot(X,
Y,
nf=0,
color="k",
lw=5,
alpha=0.7,
labels=["Sine chart", "Time (s)", "Voltage(V)"],
gl.tradingBarChart(
Cartera.get_timeData(symbolIDs[symbol_ID_indx2], periods[0]),
ax=ax2,
dataTransform=dataTransform,
AxesStyle="Normal - No xaxis",
color="k",
labels=[
"", "", symbolIDs[symbol_ID_indx2] + "(" + str(periods[0]) + "M)"
])
ax3 = gl.subplot2grid((4, 1), (2, 0), rowspan=1, colspan=1)
gl.stem(dates,
ret1,
ax=ax3,
dataTransform=dataTransform,
AxesStyle="Normal",
labels=[
"", "",
symbolIDs[symbol_ID_indx1] + "(" + str(periods[0]) + "M)"
],
legend=["Return"])
ax4 = gl.subplot2grid((4, 1), (3, 0),
rowspan=1,
colspan=1,
sharex=ax1,
sharey=ax3)
gl.stem(dates,
ret2,
ax=ax4,
dataTransform=dataTransform,
AxesStyle="Normal",
labels=["Crossing MA Strategy", "", "Price"],
nf=1)
gl.plot(dates, EMAslow, legend=["Slow"], lw=3)
gl.plot(dates, EMAfast, legend=["fast"])
ax2 = gl.plot(dates,
ul.scale(EMAfast - EMAslow),
legend=["Difference"],
nf=1,
sharex=ax1,
labels=["", "", "Events"],
fill=1,
alpha=0.3)
gl.stem(
dates,
crosses,
legend=["TradeSignal"],
)
gl.scatter(dates, ul.scale(EMAfast - EMAslow), lw=0.5, alpha=0.5)
gl.plot(dates, np.zeros(crosses.shape))
gl.subplots_adjust(left=.09,
bottom=.10,
right=.90,
top=.95,
wspace=.10,
hspace=0.01)
trade_events = myEstrategia.get_TradeEvents()
# We can observe the apparent delay !
dates = Cartera.symbols[symbol1].TDs[period1].get_dates()
########## Plotting ! ####################
gl.set_subplots(4,1)
## Plot both prices and EMAs
ax1 = gl.plot(dates,price, legend = ["Price"],
labels = ["Crossing MA Strategy", "","Price"], alpha = 0, nf = 1)
title = "Bar Chart. " + str(symbols[0]) + "(" + ul.period_dic[timeData.period]+ ")"
timeData = Cartera.symbols[]
gl.tradingBarChart(timeData, ax = ax1, legend = ["Close price"], color = "k",
labels = [title,"",r"Price ($\$$)"], AxesStyle = "Normal - No xaxis")
gl.plot(dates,EMAslow, legend = ["Slow"], lw = 1, color = "b")
gl.plot(dates,EMAfast, legend = ["fast"], lw = 1, color = "r", xaxis_mode = "hidden")
## Entry Axes
ax2 = gl.plot(dates, ul.scale(EMAfast - EMAslow), legend = ["Difference"]
,nf = 1, sharex = ax1, labels = ["","","Events"], fill = 1, alpha = 0.3)
gl.stem(dates,crosses, legend = ["TradeSignal"] )
gl.scatter(dates, ul.scale(EMAfast - EMAslow), lw = 0.5, alpha = 0.5)
gl.plot(dates,np.zeros(crosses.shape))
# We can observe the apparent delay !
#crosses, dates = Hitler.RobustXingAverages(symbols[0], Ls_list,Ll_list) #
symbolID = symbolIDs_pf[i]
myTimeData = myPortfolio.get_symbols([symbolID])[0].get_timeData(period)
returns = myTimeData.get_timeSeriesReturn(["Close"])
dates = myTimeData.get_dates()
AxesStyle = " - No xaxis"
if (i == len(symbolIDs_pf) -1):
AxesStyle = ""
if (i == 0):
title = "Bar Chart. " + str(symbolIDs) + r" . Price ($\$$)"
title2 = "Return"
else:
title = ""
title2 = ""
ylabel = symbolID + " (" + ul.period_dic[myTimeData.period] + ")"
ax = gl.tradingBarChart(myTimeData, legend = ["Close price"], color = "k",
nf = 1, sharex = axeshare, labels = [title,"",ylabel],
AxesStyle = "Normal" + AxesStyle)
# dataTransform = dataTransform)
ax = gl.stem(dates, returns, legend = ["Return"], color = "k",
nf = 1, sharex = axeshare, labels = [title2,"",""],
AxesStyle = "Normal" + AxesStyle + " - No yaxis")
# dataTransform = dataTransform)
axeshare = ax
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.10, hspace=0)
image_name = "differentSymbols.png"
gl.savefig(folder_images + image_name,
dpi = 100, sizeInches = [30, 12])
volume = timeData.get_timeSeries(["Volume"]);
Returns = timeData.get_timeSeriesReturn(seriesNames = ["Close"]);
CumReturns = timeData.get_timeSeriesCumReturn(seriesNames = ["Close"]);
nSMA = 10
SMA = timeData.SMA(n = nSMA)
ax1 = gl.plot(dates, SMA, labels = [timeData.symbolID + str(timeData.period), "Time", "Price"],
legend = ["SMA(%i)" % nSMA], nf = 1, dataTransform = dataTransform,
AxesStyle = "Normal - No xaxis", color = "cobalt blue")
gl.barchart(dates, dataHLOC, lw = 2, dataTransform = dataTransform, color = "k",
AxesStyle = "Normal - No xaxis")
# Notice the Weekends and the displacement between bars and step
############# 2: Volume
gl.stem(dates, volume, sharex = ax1, labels = [timeData.symbolID + str(timeData.period), "Time", "Volume"],
legend = ["Volume"], nf = 1, alpha = 0.5, dataTransform = dataTransform,
AxesStyle = "Normal - No xaxis")
############# 3: Returns
gl.stem(dates, Returns, sharex = ax1, labels = [timeData.symbolID + str(timeData.period), "Time", "Return"],
legend = ["Return"], nf = 1, dataTransform = dataTransform,
AxesStyle = "Normal - No xaxis")
############# 4: Commulative Returns
seriesNames = ["Close"]
gl.plot(dates, CumReturns, sharex = ax1,labels = [timeData.symbolID + str(timeData.period), "Time", "Cum Return"],
legend = ["Cum Return"], nf = 1, dataTransform = dataTransform,
AxesStyle = "Normal")
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.20, hspace=0)
ax4 = gl.subplot2grid((4, 1), (3, 0), rowspan=1, colspan=1, sharex=ax1)
## Ax1 = Close price at the end of the sessions
gl.plot(days_keys,
C,
ax=ax1,
labels=["Results " + key_classifier, "", "Close Rate"],
AxesStyle="Normal - No xaxis",
legend=["Close Rate"])
## Ax2 = 1 if the stock has gone up, zero if it has gone down
gl.stem(dates_train,
Ytrain_reg,
ax=ax2,
labels=["", "", "Target_reg"],
bottom=0.0,
AxesStyle="Normal - No xaxis",
alpha=alpha_stem,
marker=marker_stem,
color=train_color,
legend=["tr"])
gl.stem(dates_test,
Ytest_reg,
ax=ax2,
labels=["", "", "Target_reg"],
bottom=0.0,
AxesStyle="Normal - No xaxis",
color=test_color,
alpha=alpha_stem,
marker=marker_stem,
legend=["tst"])
symbolID = symbolIDs_pf[i]
myTimeData = myPortfolio.get_symbols([symbolID])[0].get_timeData(period)
returns = myTimeData.get_timeSeriesReturn(["Close"])
dates = myTimeData.get_dates()
AxesStyle = " - No xaxis"
if (i == len(symbolIDs_pf) -1):
AxesStyle = ""
if (i == 0):
title = "Bar Chart. " + str(symbolIDs) + r" . Price ($\$$)"
title2 = "Return"
else:
title = ""
title2 = ""
ylabel = symbolID + " (" + ul.period_dic[myTimeData.period] + ")"
ax = gl.tradingBarChart(myTimeData, legend = ["Close price"], color = "k",
nf = 1, sharex = axeshare, labels = [title,"",ylabel],
AxesStyle = "Normal" + AxesStyle)
# dataTransform = dataTransform)
ax = gl.stem(dates, returns, legend = ["Return"], color = "k",
nf = 1, sharex = axeshare, labels = [title2,"",""],
AxesStyle = "Normal" + AxesStyle + " - No yaxis")
# dataTransform = dataTransform)
axeshare = ax
gl.subplots_adjust(left=.09, bottom=.10, right=.90, top=.95, wspace=.10, hspace=0)
image_name = "differentSymbols.png"
gl.savefig(folder_images + image_name,
dpi = 100, sizeInches = [30, 12])
