experiments for report

.gitmodules

@@ -0,0 +1,3 @@
+[submodule "result_path"]
+	path = result_path
+	url = gitea@gitea-s2i2s.isti.cnr.it:moreo/result_table.git

Census/simulated-cnr/commons.py

@@ -0,0 +1,22 @@
+from dataclasses import dataclass
+
+
+@dataclass
+class DataConfig:
+    n_features: int
+    n_informative: int
+    n_redundant: int
+    n_clusters_per_class: int
+    flip_y:float
+
+config_easy     = DataConfig(n_features=2, n_informative=2, n_redundant=0, n_clusters_per_class=1, flip_y=0.0)
+config_medium   = DataConfig(n_features=10, n_informative=5, n_redundant=2, n_clusters_per_class=2, flip_y=0.01)
+config_hard     = DataConfig(n_features=50, n_informative=20, n_redundant=10, n_clusters_per_class=4, flip_y=0.05)
+config_veryhard = DataConfig(n_features=100, n_informative=30, n_redundant=20, n_clusters_per_class=8, flip_y=0.05)
+
+configs = [
+    config_easy,
+    config_medium,
+    config_hard,
+    config_veryhard
+]

Census/simulated-cnr/gen_data.py

@@ -6,17 +6,20 @@ from quapy.data import LabelledCollection
 from quapy.protocol import UniformPrevalenceProtocol
 import quapy.functional as F
 import pandas as pd
+from commons import configs
 
 random_state = 0
 
-n_features = 10
+
 n_areas = 50
 n_per_area = 1_000
 population_size = n_areas * n_per_area
 n_experiments = 100
 n_survey = population_size//n_experiments
 
-print(f'{n_features=}')
+
+for config in configs:
+    print(f'{config.n_features=}')
     print(f'{n_areas=}')
     print(f'{n_per_area=}')
     print(f'{population_size=}')
@@ -25,14 +28,14 @@ print(f'{n_survey=}')
 
     X, y = make_classification(
         n_samples=population_size * 100,
-    n_features=n_features,
-    n_informative=n_features//2,
-    n_redundant=2,
+        n_features=config.n_features,
+        n_informative=config.n_informative,
+        n_redundant=config.n_redundant,
         n_repeated=0,
         n_classes=2,
-    n_clusters_per_class=2,
+        n_clusters_per_class=config.n_clusters_per_class,
         weights=[0.5, 0.5],
-    flip_y=0.01,
+        flip_y=config.flip_y,
         class_sep=1.0,
         hypercube=True,
         shift=0.0,
@@ -73,7 +76,7 @@ data_dic = {
         'ID': idx,
         'Y': data_y,
     }
-for feat_id in range(n_features):
+    for feat_id in range(config.n_features):
         data_dic[f'X_{feat_id}'] = data_X[:,feat_id]
     data_dic['area'] = data_area
 
@@ -82,7 +85,7 @@ for experiment_id, experiment_selection in enumerate(experiment_selections):
 
     df = pd.DataFrame(data_dic)
 
-data_path = f'./data/data_nF{n_features}_nA{n_areas}_P{population_size}_nExp{n_experiments}.csv'
+    data_path = f'./data/data_nF{config.n_features}_nA{n_areas}_P{population_size}_nExp{n_experiments}.csv'
     os.makedirs(Path(data_path).parent, exist_ok=True)
     df.to_csv(data_path, index=0)
 

Census/simulated-cnr/main.py

@@ -3,15 +3,14 @@ from os.path import join
 
 import numpy as np
 import pandas as pd
-from sklearn.calibration import CalibratedClassifierCV
 from sklearn.linear_model import LogisticRegression, LogisticRegressionCV
 from pathlib import Path
 from quapy.data import LabelledCollection
-from quapy.model_selection import GridSearchQ
-from quapy.protocol import APP
 from quapy.method.aggregative import PACC, PCC, EMQ, DMy, ACC, KDEyML, CC
-import quapy.functional as F
 from tqdm import tqdm
+from commons import configs
+
+from src.new_table import LatexTable
 
 pd.set_option('display.max_columns', None)
 pd.set_option('display.width', 1000)
@@ -42,13 +41,17 @@ def methods():
     yield 'PACC', PACC(classifier=LogisticRegression())
     yield 'EMQ', EMQ(classifier=LogisticRegression())
     yield 'KDEy', KDEyML(classifier=LogisticRegression(), bandwidth=0.05)
-    yield 'KDEy01', KDEyML(classifier=LogisticRegression())
+    # yield 'KDEy01', KDEyML(classifier=LogisticRegression())
 
 
-data_path = './data/data_nF10_nA50_P50000_nExp100.csv'
+for config in configs:
 
-config = Path(data_path).name.replace('.csv','')
-result_dir = f'./results/{config}'
+    print(f'Running {config}')
+
+    config_name = f'data_nF{config.n_features}_nA50_P50000_nExp100'
+    data_path = f'./data/{config_name}.csv'
+
+    result_dir = f'./results/{config_name}'
     os.makedirs(result_dir, exist_ok=True)
 
     X, y, A, numExperiments, df = load_data(data_path)
@@ -58,8 +61,18 @@ n_areas = len(areas)
 
     methods_results = []
 
-for q_name, quantifier in methods():
+    # load baseline result from UniPI
+    baseline_path = join(result_dir, 'Risultati_SAE.csv')
+    if os.path.exists(baseline_path):
+        unipi_baseline_df = pd.read_csv(baseline_path, index_col=0, sep=';')
+        unipi_baseline_df = unipi_baseline_df.rename(columns={'AE(SAE)': 'AE'})
+        unipi_baseline_name = "SAE"
+        methods_results.append(unipi_baseline_df)
+    else:
+        unipi_baseline_name = None
 
+    # run quantification methods
+    for q_name, quantifier in methods():
         result_path = join(result_dir, f'{q_name}.csv')
         if os.path.exists(result_path):
             method_results = pd.read_csv(result_path, index_col=0)
@@ -107,6 +120,18 @@ for q_name, quantifier in methods():
         methods_results.append(method_results)
 
     methods_results = pd.concat(methods_results)
+
+    methods_results["area"] = methods_results["area"].astype(str).str.zfill(2)
+    latex_table = LatexTable.from_dataframe(methods_results, method='method', benchmark='area', value='AE')
+    latex_table.format.configuration.resizebox=True
+
+    methods_order = [m for m, _ in methods()]
+    if unipi_baseline_name is not None:
+        methods_order = [unipi_baseline_name] + methods_order
+
+    latex_table.reorder_methods(methods_order)
+    latex_table.latexPDF(pdf_path=join('./tables', f'{config_name}.pdf'), tabular_dir=f'tabular_{config_name}', landscape=False)
+
     pv = methods_results.pivot_table(
         index='area',
         columns='method',

result_path

@@ -0,0 +1 @@
+Subproject commit 816a4c675e2919ea0ec4dd2ba9bf0d518d53dc17