From 9a90c4d4e473f3d7f155afef166ccb1ff4d05249 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Petr=C3=B4nio=20C=C3=A2ndido=20de=20Lima=20e=20Silva?= Date: Sat, 21 Jan 2017 00:38:32 -0200 Subject: [PATCH] Novos benchmarks --- benchmarks/benchmarks.py | 130 ++++++++++++++------------------------- fts.py | 2 +- partitioners/Util.py | 23 ++++--- pfts.py | 4 +- 4 files changed, 62 insertions(+), 97 deletions(-) diff --git a/benchmarks/benchmarks.py b/benchmarks/benchmarks.py index a2d83d2..f939aa6 100644 --- a/benchmarks/benchmarks.py +++ b/benchmarks/benchmarks.py @@ -35,32 +35,32 @@ def plotDistribution(dist): vmin=0, vmax=1, edgecolors=None) -def plotComparedSeries(original, models, colors, typeonlegend=False, save=False, file=None,tam=[20, 5]): +def plotComparedSeries(original, models, colors, typeonlegend=False, save=False, file=None,tam=[20, 5],intervals=True): fig = plt.figure(figsize=tam) ax = fig.add_subplot(111) mi = [] ma = [] - ax.plot(original, color='black', label="Original") + ax.plot(original, color='black', label="Original",linewidth=1.5) count = 0 for fts in models: if fts.hasPointForecasting: forecasted = fts.forecast(original) - mi.append(min(forecasted)) - ma.append(max(forecasted)) + mi.append(min(forecasted)*0.95) + ma.append(max(forecasted)*1.05) for k in np.arange(0, fts.order): forecasted.insert(0, None) lbl = fts.shortname if typeonlegend: lbl += " (Point)" ax.plot(forecasted, color=colors[count], label=lbl, ls="-") - if fts.hasIntervalForecasting: + if fts.hasIntervalForecasting and intervals: forecasted = fts.forecastInterval(original) lower = [kk[0] for kk in forecasted] upper = [kk[1] for kk in forecasted] - mi.append(min(lower)) - ma.append(max(upper)) + mi.append(min(lower)*0.95) + ma.append(max(upper)*1.05) for k in np.arange(0, fts.order): lower.insert(0, None) upper.insert(0, None) @@ -430,101 +430,61 @@ def compareModelsTable(original, models_fo, models_ho): return sup + header + body + "\\end{tabular}" -from pyFTS import hwang - - -def HOSelecaoSimples_MenorRMSE(original, parameters, orders): +def simpleSearch_RMSE(original, model, partitions, orders, save=False, file=None,tam=[10, 15],plotforecasts=False,elev=30, azim=144): ret = [] - errors = np.array([[0 for k in range(len(parameters))] for kk in range(len(orders))]) + errors = np.array([[0 for k in range(len(partitions))] for kk in range(len(orders))]) forecasted_best = [] - print("Série Original") - fig = plt.figure(figsize=[20, 12]) - fig.suptitle("Comparação de modelos ") - ax0 = fig.add_axes([0, 0.5, 0.6, 0.45]) # left, bottom, width, height - ax0.set_xlim([0, len(original)]) - ax0.set_ylim([min(original), max(original)]) - ax0.set_title('Série Temporal') - ax0.set_ylabel('F(T)') - ax0.set_xlabel('T') - ax0.plot(original, label="Original") - min_rmse = 100000.0 + fig = plt.figure(figsize=tam) + #fig.suptitle("Comparação de modelos ") + if plotforecasts: + ax0 = fig.add_axes([0, 0.5, 0.9, 0.45]) # left, bottom, width, height + ax0.set_xlim([0, len(original)]) + ax0.set_ylim([min(original)*0.9, max(original)*1.1]) + ax0.set_title('Forecasts') + ax0.set_ylabel('F(T)') + ax0.set_xlabel('T') + min_rmse = 1000000.0 best = None pc = 0 - for p in parameters: + for p in partitions: oc = 0 for o in orders: sets = Grid.GridPartitionerTrimf(original, p) - fts = hwang.HighOrderFTS(o, "k = " + str(p) + " w = " + str(o)) - fts.train(original, sets) - forecasted = [fts.forecast(original, xx) for xx in range(o, len(original))] - error = Measures.rmse(np.array(forecasted), np.array(original[o:])) + fts = model("q = " + str(p) + " n = " + str(o)) + fts.train(original, sets, o) + forecasted = fts.forecast(original) + error = Measures.rmse(np.array(original[o:]),np.array(forecasted[:-1])) + mape = Measures.mape(np.array(original[o:]), np.array(forecasted[:-1])) + #print(original[o:]) + #print(forecasted[-1]) for kk in range(o): forecasted.insert(0, None) - ax0.plot(forecasted, label=fts.name) - print(o, p, error) + if plotforecasts: ax0.plot(forecasted, label=fts.name) + #print(o, p, mape) errors[oc, pc] = error if error < min_rmse: min_rmse = error best = fts forecasted_best = forecasted - oc = oc + 1 - pc = pc + 1 - handles0, labels0 = ax0.get_legend_handles_labels() - ax0.legend(handles0, labels0) - ax1 = Axes3D(fig, rect=[0.6, 0.5, 0.45, 0.45], elev=30, azim=144) + oc += 1 + pc += 1 + #print(min_rmse) + if plotforecasts: + #handles0, labels0 = ax0.get_legend_handles_labels() + #ax0.legend(handles0, labels0) + ax0.plot(original, label="Original", linewidth=3.0, color="black") + ax1 = Axes3D(fig, rect=[0, 1, 0.9, 0.45], elev=elev, azim=azim) + if not plotforecasts: ax1 = Axes3D(fig, rect=[0, 1, 0.9, 0.9], elev=elev, azim=azim) # ax1 = fig.add_axes([0.6, 0.5, 0.45, 0.45], projection='3d') - ax1.set_title('Comparação dos Erros Quadráticos Médios por tamanho da janela') - ax1.set_ylabel('RMSE') - ax1.set_xlabel('Quantidade de Partições') - ax1.set_zlabel('W') - X, Y = np.meshgrid(parameters, orders) + ax1.set_title('Error Surface') + ax1.set_ylabel('Model order') + ax1.set_xlabel('Number of partitions') + ax1.set_zlabel('RMSE') + X, Y = np.meshgrid(partitions, orders) surf = ax1.plot_surface(X, Y, errors, rstride=1, cstride=1, antialiased=True) ret.append(best) ret.append(forecasted_best) - # Modelo diferencial - print("\nSérie Diferencial") - errors = np.array([[0 for k in range(len(parameters))] for kk in range(len(orders))]) - forecastedd_best = [] - ax2 = fig.add_axes([0, 0, 0.6, 0.45]) # left, bottom, width, height - ax2.set_xlim([0, len(original)]) - ax2.set_ylim([min(original), max(original)]) - ax2.set_title('Série Temporal') - ax2.set_ylabel('F(T)') - ax2.set_xlabel('T') - ax2.plot(original, label="Original") - min_rmse = 100000.0 - bestd = None - pc = 0 - for p in parameters: - oc = 0 - for o in orders: - sets = Grid.GridPartitionerTrimf(Transformations.differential(original), p) - fts = hwang.HighOrderFTS(o, "k = " + str(p) + " w = " + str(o)) - fts.train(original, sets) - forecasted = [fts.forecastDiff(original, xx) for xx in range(o, len(original))] - error = Measures.rmse(np.array(forecasted), np.array(original[o:])) - for kk in range(o): - forecasted.insert(0, None) - ax2.plot(forecasted, label=fts.name) - print(o, p, error) - errors[oc, pc] = error - if error < min_rmse: - min_rmse = error - bestd = fts - forecastedd_best = forecasted - oc = oc + 1 - pc = pc + 1 - handles0, labels0 = ax2.get_legend_handles_labels() - ax2.legend(handles0, labels0) - ax3 = Axes3D(fig, rect=[0.6, 0.0, 0.45, 0.45], elev=30, azim=144) - # ax3 = fig.add_axes([0.6, 0.0, 0.45, 0.45], projection='3d') - ax3.set_title('Comparação dos Erros Quadráticos Médios') - ax3.set_ylabel('RMSE') - ax3.set_xlabel('Quantidade de Partições') - ax3.set_zlabel('W') - X, Y = np.meshgrid(parameters, orders) - surf = ax3.plot_surface(X, Y, errors, rstride=1, cstride=1, antialiased=True) - ret.append(bestd) - ret.append(forecastedd_best) + Util.showAndSaveImage(fig,file,save) + return ret diff --git a/fts.py b/fts.py index 1ede686..67dabd1 100644 --- a/fts.py +++ b/fts.py @@ -45,7 +45,7 @@ class FTS: def forecastAheadDistribution(self, data, steps): pass - def train(self, data, sets): + def train(self, data, sets, order=1): pass def getMidpoints(self, flrg): diff --git a/partitioners/Util.py b/partitioners/Util.py index 3c4bcf8..bff9ce4 100644 --- a/partitioners/Util.py +++ b/partitioners/Util.py @@ -9,21 +9,26 @@ from pyFTS.common import Membership, Util def plotSets(data, sets, titles, tam=[12, 10], save=False, file=None): num = len(sets) - fig = plt.figure(figsize=tam) + #fig = plt.figure(figsize=tam) maxx = max(data) minx = min(data) - h = 1/num - for k in range(num): - ax0 = fig.add_axes([0, (k+1)*h, 0.65, h*0.7]) # left, bottom, width, height - ax0.set_title(titles[k]) - ax0.set_ylim([0, 1]) - ax0.set_xlim([minx, maxx]) + #h = 1/num + #print(h) + fig, axes = plt.subplots(nrows=num, ncols=1,figsize=tam) + for k in np.arange(0,num): + #ax = fig.add_axes([0.05, 1-(k*h), 0.9, h*0.7]) # left, bottom, width, height + ax = axes[k] + ax.set_title(titles[k]) + ax.set_ylim([0, 1]) + ax.set_xlim([minx, maxx]) for s in sets[k]: if s.mf == Membership.trimf: - ax0.plot([s.parameters[0],s.parameters[1],s.parameters[2]],[0,1,0]) + ax.plot([s.parameters[0],s.parameters[1],s.parameters[2]],[0,1,0]) elif s.mf == Membership.gaussmf: tmpx = [ kk for kk in np.arange(s.lower, s.upper)] tmpy = [s.membership(kk) for kk in np.arange(s.lower, s.upper)] - ax0.plot(tmpx, tmpy) + ax.plot(tmpx, tmpy) + + plt.tight_layout() Util.showAndSaveImage(fig, file, save) \ No newline at end of file diff --git a/pfts.py b/pfts.py index 09f8362..1e80105 100644 --- a/pfts.py +++ b/pfts.py @@ -37,8 +37,8 @@ class ProbabilisticFLRG(hofts.HighOrderFLRG): class ProbabilisticFTS(ifts.IntervalFTS): def __init__(self, name): - super(ProbabilisticFTS, self).__init__("PIFTS") - self.shortname = "PIFTS " + name + super(ProbabilisticFTS, self).__init__("PFTS") + self.shortname = "PFTS " + name self.name = "Probabilistic FTS" self.detail = "Silva, P.; Guimarães, F.; Sadaei, H." self.flrgs = {}