Update abalone.py
Browse files- abalone.py +17 -2
abalone.py
CHANGED
@@ -58,6 +58,17 @@ features_types_per_config = {
|
|
58 |
"viscera_weight": datasets.Value("float64"),
|
59 |
"shell_weight": datasets.Value("float64"),
|
60 |
"number_of_rings": datasets.Value("int8")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
61 |
}
|
62 |
}
|
63 |
features_per_config = {k: datasets.Features(features_types_per_config[k]) for k in features_types_per_config}
|
@@ -73,8 +84,8 @@ class Abalone(datasets.GeneratorBasedBuilder):
|
|
73 |
# dataset versions
|
74 |
DEFAULT_CONFIG = "abalone"
|
75 |
BUILDER_CONFIGS = [
|
76 |
-
AbaloneConfig(name="abalone",
|
77 |
-
|
78 |
]
|
79 |
|
80 |
|
@@ -95,6 +106,10 @@ class Abalone(datasets.GeneratorBasedBuilder):
|
|
95 |
data = pandas.read_csv(filepath, header=None)
|
96 |
data.columns = _BASE_FEATURE_NAMES
|
97 |
|
|
|
|
|
|
|
|
|
98 |
for row_id, row in data.iterrows():
|
99 |
data_row = dict(row)
|
100 |
|
|
|
58 |
"viscera_weight": datasets.Value("float64"),
|
59 |
"shell_weight": datasets.Value("float64"),
|
60 |
"number_of_rings": datasets.Value("int8")
|
61 |
+
},
|
62 |
+
"binary": {
|
63 |
+
"sex": datasets.Value("string"),
|
64 |
+
"length": datasets.Value("float64"),
|
65 |
+
"diameter": datasets.Value("float64"),
|
66 |
+
"height": datasets.Value("float64"),
|
67 |
+
"whole_weight": datasets.Value("float64"),
|
68 |
+
"shucked_weight": datasets.Value("float64"),
|
69 |
+
"viscera_weight": datasets.Value("float64"),
|
70 |
+
"shell_weight": datasets.Value("float64"),
|
71 |
+
"is_old": datasets.ClassLabel(num_classes=2)
|
72 |
}
|
73 |
}
|
74 |
features_per_config = {k: datasets.Features(features_types_per_config[k]) for k in features_types_per_config}
|
|
|
84 |
# dataset versions
|
85 |
DEFAULT_CONFIG = "abalone"
|
86 |
BUILDER_CONFIGS = [
|
87 |
+
AbaloneConfig(name="abalone", description="Abalone for regression."),
|
88 |
+
AbaloneConfig(name="binary", description="Abalone for binary classification."),
|
89 |
]
|
90 |
|
91 |
|
|
|
106 |
data = pandas.read_csv(filepath, header=None)
|
107 |
data.columns = _BASE_FEATURE_NAMES
|
108 |
|
109 |
+
if self.config.name == "binary":
|
110 |
+
data = data.rename(columns={"number_of_rings": "is_old"})
|
111 |
+
data["is_old"] = data["is_old"].apply(lambda x: 1 if x > 9 else 0)
|
112 |
+
|
113 |
for row_id, row in data.iterrows():
|
114 |
data_row = dict(row)
|
115 |
|