GOPSA/plot_simulated_expe.py at main · apmellot/GOPSA · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
# %%
from pathlib import Path
import matplotlib.pyplot as plt
import pandas as pd
import seaborn as sns
import numpy as np

from sklearn.metrics import r2_score


DEBUG = False

# activate latex text rendering
plt.rc('text', usetex=True)

if DEBUG:
    all_results = pd.read_csv(
        './results/simulated_expe_debug.csv',
        index_col=0
    )
else:
    all_results = pd.read_csv(
        './results/simulated_expe.csv',
        index_col=0
    )

all_results.reset_index(inplace=True, drop=True)
FIGURES_FOLDER = Path('./figures')
FIGURES_FOLDER.mkdir(parents=True, exist_ok=True)

all_results = all_results.replace(
    [
        'shift_X',
        'shift_Y',
        'shift_XY',
    ],
    [
        'Shift in ' + r'$X$',
        'Shift in ' + r'$y$',
        'Shift in ' + r'($X$, $y$)',
    ]
)

all_results = all_results.replace(
    [
        'dummy',
        'baseline_no_recenter',
        'baseline_green',
        'baseline_recenter',
        'baseline_rescale',
        'baseline_fit_intercept',
        'geodesic_optim'
    ],
    [
        r'\texttt{DO Dummy}',
        r'\texttt{No DA}',
        r'\texttt{GREEN}',
        r'\texttt{Re-center}',
        r'\texttt{Re-scale}',
        r'\texttt{DO Intercept}',
        r'\texttt{GOPSA}',
    ]
)
order = [
    r'\texttt{DO Dummy}',
    r'\texttt{No DA}',
    r'\texttt{GREEN}',
    r'\texttt{Re-center}',
    r'\texttt{Re-scale}',
    r'\texttt{DO Intercept}',
    r'\texttt{GOPSA}'
]

# group by (method, scenario, parameter, random_state)
# and compute R2 from y_true and y_pred
all_results = all_results.groupby(
    ['method', 'scenario', 'parameter', 'random_state']
).apply(
    lambda x: pd.Series({
        'r2': r2_score(x['y_true'], x['y_pred'])
    })
).reset_index()

sns.set_theme(style="ticks", font_scale=2)
sns.set_palette('colorblind')
palette = np.array(sns.color_palette("colorblind"))[:7]
palette = list(palette[[3, 1, 5, 2, 6, 0, 4]])
g = sns.FacetGrid(
    all_results,
    col="scenario",
    col_wrap=3,
    legend_out=True,
    hue='method',
    hue_kws=dict(marker=['s', 'o', '<', 'X', 'v', '>', 'P'],
                 ls=['--', '-.', ':',
                     (0, (5, 10)),
                     (0, (3, 5, 1, 5, 1, 5)),
                     (0, (1, 10)),
                     '-']),
    height=4, aspect=1,
    margin_titles=True,
    sharex=False,
    hue_order=order,
    palette=palette
)

g.map_dataframe(sns.lineplot, 'parameter', 'r2', markersize=10)
g.set(ylim=(-0.15, 1.1))
g.set_axis_labels("Parameter value", r"$R^2$")

# g.set(xscale='log')
g.add_legend(title='Methods')
g.set_titles(col_template='{col_name}')
letters = ['A', 'B', 'C', 'D']
for i, ax in enumerate(g.axes.flat):
    ax.set_xlabel(r'$\xi$')

    # Change x_ticks with "No shift", "Max. shift"
    shift_values = all_results[
        all_results['scenario'] == all_results['scenario'].unique()[i]
    ]['parameter'].unique()
    min_shift = min(shift_values)
    max_shift = max(shift_values)
    ax.set_xticks([min_shift, max_shift])
    ax.set_xticklabels(['No shift', 'Max. shift'])

    # Add letter
    ax.annotate(
        text=letters[i], xy=(-0.1, 1.13),
        xycoords=('axes fraction'), fontsize=25,
        weight='bold'
    )

sns.despine(trim=True)
g.tight_layout()

if DEBUG:
    g.savefig(
        FIGURES_FOLDER / "simulated_expe_debug.pdf",
        bbox_inches='tight'
    )
else:
    g.savefig(
        FIGURES_FOLDER / "simulated_expe.pdf",
        bbox_inches='tight'
    )