def make_plot(*data_files):
table = []
for f in data_files:
t = np.loadtxt(f, delimiter=',')
t[:, 0] = np.array([int(D) for D in t[:, 0]])
table.append(t)
table = np.concatenate(table)
## customize plot below ##
p = plot.ErrorBar()
p.xlabel = 'number of dimensions'
p.legends = [
r'$\mathcal{E}^{H}$', r'$\mathcal{E}^{L}$', r'$\mathcal{E}^{S}$',
r'$k$-means', r'GMM'
]
p.colors = ['b', 'r', 'g', 'm', 'c']
p.symbols = ['o', 's', 'D', '^', 'v']
p.lines = ['-', '-', '-', '-', '-']
#p.output = './experiments_figs/normal_highdim_mean.pdf'
p.output = './experiments_figs/normal_highdim_cov.pdf'
#p.bayes = 0.86
p.bayes = 0.9537075
#p.xlim = [10, 200]
p.xlim = [10, 700]
p.make_plot(table)
def make_plot_difference(*data_files):
table = []
for f in data_files:
t = np.loadtxt(f, delimiter=',')
t[:,0] = np.array([int(D) for D in t[:,0]])
table.append(t)
table = np.concatenate(table)
## customize plot below ##
p = plot.ErrorBar()
p.xlabel = 'number of points'
p.ylabel = 'difference in accuracy'
p.legends = [
r'$\mathcal{E}^{H} - \mathcal{E}^{L}$',
r'$\mathcal{E}^{H} - \mathcal{E}^{S}$',
]
p.colors = ['b', 'r']
p.symbols = ['o', 's']
#p.output = './experiments_figs/normal_kernels_difference.pdf'
p.output = './experiments_figs/lognormal_kernels_difference.pdf'
p.doublex = True
p.legcols = 1
#p.bayes = 0.0
p.xlim = [10, 400]
p.make_plot(table)
def make_plot(*data_files):
table = []
for f in data_files:
t = np.loadtxt(f, delimiter=',')
t[:,0] = np.array([int(D) for D in t[:,0]])
table.append(t)
table = np.concatenate(table)
## customize plot below ##
p = plot.ErrorBar()
p.xlabel = 'number of points'
p.legends = [r'$\mathcal{E}^{H}$, $\rho$',
r'$\mathcal{E}^{H}$, $\rho_{1/2}$',
r'$\mathcal{E}^{H}$, $\rho_{e}$',
r'$k$-means',
r'GMM']
p.colors = ['b', 'r', 'g', 'm', 'c']
p.symbols = ['o', 's', 'D', '^', 'v']
p.lines = ['-', '-', '-', '-', '-']
p.output = './experiments_figs/normal_kernels.pdf'
#p.output = './experiments_figs/lognormal_kernels.pdf'
p.doublex = True
p.bayes = 0.9
p.xlim = [10, 400]
p.make_plot(table)
def make_plot(*data_files):
table = []
for f in data_files:
t = np.loadtxt(f, delimiter=',')
t[:, 0] = np.array([int(D) for D in t[:, 0]])
table.append(t)
table = np.concatenate(table)
## customize plot below ##
p = plot.ErrorBar()
p.xlabel = r'$n$'
p.ylabel = r'accuracy'
p.legends = [r'$\mathcal{E}^{H}$-clustering', r'$k$-means', 'GMM']
p.colors = ['b', 'r', 'g']
p.lines = ['-', '-', '-']
#p.output = './experiments_figs2/1D_normal.pdf'
p.output = './experiments_figs2/1D_lognormal.pdf'
p.bayes = 0.88
#p.bayes = 0.852
p.xlim = [10, 800]
#p.ylim = [0.75,0.89]
p.ylim = [0.5, 0.89]
#p.loc = [0.45,0.5]
p.loc = [0.45, 0.3]
p.make_plot(table)
def make_plot_difference(*data_files):
table = []
for f in data_files:
t = np.loadtxt(f, delimiter=',')
t[:, 0] = np.array([int(D) for D in t[:, 0]])
table.append(t)
table = np.concatenate(table)
## customize plot below ##
p = plot.ErrorBar()
p.xlabel = r'$\#$ points'
p.ylabel = 'difference in accuracy'
p.legends = [
#r'$\mathcal{E}^{H} - \textnormal{kernel $k$-means}$',
#r'$\mathcal{E}^{H} - \textnormal{spectral-clustering}$',
r'$\mathcal{E}^{H}\mbox{-clustering} - \mbox{kernel $k$-means}$',
r'$\mathcal{E}^{H}\mbox{-clustering} - \mbox{spectral clustering}$',
]
p.colors = ['b', 'r']
p.symbols = ['o', 's']
p.output = './experiments_figs2/normal_kernels_difference.pdf'
#p.output = './experiments_figs2/lognormal_kernels_difference.pdf'
#p.doublex = True
p.legcols = 1
#p.bayes = 0.0
p.xlim = [10, 400]
#p.loc = 1
p.loc = 3
p.make_plot(table)
def make_plot(*data_files):
table = []
for f in data_files:
t = np.loadtxt(f, delimiter=',')
t[:, 0] = np.array([int(D) for D in t[:, 0]])
table.append(t)
table = np.concatenate(table)
## customize plot below ##
p = plot.ErrorBar()
p.xlabel = r'$\#$ points'
p.ylabel = 'accuracy'
p.legends = [
r'$\mathcal{E}^{H}$-clustering, $\rho_1$',
r'$\mathcal{E}^{H}$-clustering, $\rho_{1/2}$',
r'$\mathcal{E}^{H}$-clustering, $\widetilde{\rho}_{1}$', r'$k$-means',
r'GMM'
]
p.colors = ['b', 'r', 'g', 'm', 'c']
p.symbols = ['o', 's', 'D', '^', 'v']
p.lines = ['-', '-', '-', '-', '-']
#p.output = './experiments_figs2/normal_kernels.pdf'
p.output = './experiments_figs2/lognormal_kernels.pdf'
#p.doublex = True
p.bayes = 0.9
p.xlim = [10, 400]
#p.ylim = [0.6, 0.91]
p.ylim = [0.55, 0.91]
p.loc = 0
p.make_plot(table)
def make_plot(*data_files):
table = []
for f in data_files:
t = np.loadtxt(f, delimiter=',')
t[:, 0] = np.array([int(D) for D in t[:, 0]])
table.append(t)
table = np.concatenate(table)
## customize plot below ##
p = plot.ErrorBar()
p.xlabel = r'$\#$ dimensions'
p.ylabel = r'accuracy'
p.legends = [
r'$\mathcal{E}^{H}$-clustering',
r'kernel $k$-means',
#r'$\mathcal{E}^L$',
r'spectral clustering',
r'$k$-means',
r'GMM'
]
p.loc = 3
p.colors = ['b', 'r', 'g', 'm', 'c']
p.symbols = ['o', 's', 'D', '^', 'v']
p.lines = ['-', '-', '-', '-', '-']
p.output = './experiments_figs2/normal_highdim_mean.pdf'
#p.output = './experiments_figs2/normal_highdim_cov.pdf'
p.bayes = 0.86
#p.bayes = 0.9537075
p.xlim = [10, 200]
#p.xlim = [10, 700]
p.ylim = [0.55, 0.87]
#p.ylim = [0.55, 0.965]
p.make_plot(table)
def make_plot(*data_files):
table = []
for f in data_files:
t = np.loadtxt(f, delimiter=',')
t[:,0] = np.array([int(D) for D in t[:,0]])
table.append(t)
table = np.concatenate(table)
## customize plot below ##
p = plot.ErrorBar()
p.xlabel = 'number of unbalanced points'
p.legends = [r'$\mathcal{E}^{H}$',
r'$\mathcal{E}^{L}$',
r'$\mathcal{E}^{S}$',
r'$k$-means',
r'GMM']
p.colors = ['b', 'r', 'g', 'm', 'c']
p.symbols = ['o', 's', 'D', '^', 'v']
p.xlim = [0,240]
p.output = './experiments_figs/normal_unbalanced.pdf'
p.make_plot(table)
