Doctorado-ML/benchmark
1
0
Fork 0
mirror of https://github.com/Doctorado-ML/benchmark.git synced 2025-08-16 16:05:54 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add balance to dataset report

Browse Source
This commit is contained in:
Ricardo Montañana Gómez
2022-03-28 12:03:07 +02:00
parent 19f5df0a4c
commit b9abbac699
1 changed files with 9 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

12
src/report
View File

@@ -92,8 +92,8 @@ def default_report():
sets = Datasets() sets = Datasets()
color_line = TextColor.LINE1 color_line = TextColor.LINE1
print(color_line, end="") print(color_line, end="")
print(f"{'Dataset':30s} Samp. Feat Cls") print(f"{'Dataset':30s} Samp. Feat Cls Balance")
print("=" * 30 + " ===== ==== ===") print("=" * 30 + " ===== ==== === " + "=" * 40)
for line in sets: for line in sets:
X, y = sets.load(line) X, y = sets.load(line)
color_line = ( color_line = (
@@ -101,10 +101,16 @@ def default_report():
if color_line == TextColor.LINE1 if color_line == TextColor.LINE1
else TextColor.LINE1 else TextColor.LINE1
) )
values, counts = np.unique(y, return_counts=True)
comp = ""
sep = ""
for value, count in zip(values, counts):
comp += f"{sep}{count/sum(counts)*100:5.2f}%"
sep = "/ "
print(color_line, end="") print(color_line, end="")
print( print(
f"{line:30s} {X.shape[0]:5,d} {X.shape[1]:4d} " f"{line:30s} {X.shape[0]:5,d} {X.shape[1]:4d} "
f"{len(np.unique(y)):3d}" f"{len(np.unique(y)):3d} {comp:40s}"
) )