Datasets:
Upload heart.py
Browse files
heart.py
CHANGED
|
@@ -40,60 +40,9 @@ _CITATION = """
|
|
| 40 |
|
| 41 |
# Dataset info
|
| 42 |
urls_per_split = {
|
| 43 |
-
"cleveland": {"train": "https://huggingface.co/datasets/mstz/heart/raw/main/processed.cleveland.data"},
|
| 44 |
"hungary": {"train": "https://huggingface.co/datasets/mstz/heart/raw/main/processed.hungarian.data"},
|
| 45 |
-
"switzerland": {"train": "https://huggingface.co/datasets/mstz/heart/raw/main/processed.switzerland.data"},
|
| 46 |
-
"va": {"train": "https://huggingface.co/datasets/mstz/heart/raw/main/processed.va.data"}
|
| 47 |
}
|
| 48 |
features_types_per_config = {
|
| 49 |
-
"cleveland": {
|
| 50 |
-
"age": datasets.Value("int8"),
|
| 51 |
-
"is_male": datasets.Value("bool"),
|
| 52 |
-
"type_of_chest_pain": datasets.Value("string"),
|
| 53 |
-
"resting_blood_pressure": datasets.Value("float32"),
|
| 54 |
-
"serum_cholesterol": datasets.Value("float32"),
|
| 55 |
-
"fasting_blood_sugar": datasets.Value("float32"),
|
| 56 |
-
"rest_electrocardiographic_type": datasets.Value("string"),
|
| 57 |
-
"maximum_heart_rate": datasets.Value("float32"),
|
| 58 |
-
"has_exercise_induced_angina": datasets.Value("bool"),
|
| 59 |
-
"depression_induced_by_exercise": datasets.Value("float32"),
|
| 60 |
-
"slope_of_peak_exercise": datasets.Value("float32"),
|
| 61 |
-
"number_of_major_vessels_colored_by_flourosopy": datasets.Value("int16"),
|
| 62 |
-
"thal": datasets.Value("float32"),
|
| 63 |
-
"has_hearth_disease": datasets.ClassLabel(num_classes=2, names=("no", "yes"))
|
| 64 |
-
},
|
| 65 |
-
"va": {
|
| 66 |
-
"age": datasets.Value("int8"),
|
| 67 |
-
"is_male": datasets.Value("bool"),
|
| 68 |
-
"type_of_chest_pain": datasets.Value("string"),
|
| 69 |
-
"resting_blood_pressure": datasets.Value("float32"),
|
| 70 |
-
"serum_cholesterol": datasets.Value("float32"),
|
| 71 |
-
"fasting_blood_sugar": datasets.Value("float32"),
|
| 72 |
-
"rest_electrocardiographic_type": datasets.Value("string"),
|
| 73 |
-
"maximum_heart_rate": datasets.Value("float32"),
|
| 74 |
-
"has_exercise_induced_angina": datasets.Value("bool"),
|
| 75 |
-
"depression_induced_by_exercise": datasets.Value("float32"),
|
| 76 |
-
"slope_of_peak_exercise": datasets.Value("float32"),
|
| 77 |
-
"number_of_major_vessels_colored_by_flourosopy": datasets.Value("int16"),
|
| 78 |
-
"thal": datasets.Value("float32"),
|
| 79 |
-
"has_hearth_disease": datasets.ClassLabel(num_classes=2, names=("no", "yes"))
|
| 80 |
-
},
|
| 81 |
-
"switzerland": {
|
| 82 |
-
"age": datasets.Value("int8"),
|
| 83 |
-
"is_male": datasets.Value("bool"),
|
| 84 |
-
"type_of_chest_pain": datasets.Value("string"),
|
| 85 |
-
"resting_blood_pressure": datasets.Value("float32"),
|
| 86 |
-
"serum_cholesterol": datasets.Value("float32"),
|
| 87 |
-
"fasting_blood_sugar": datasets.Value("float32"),
|
| 88 |
-
"rest_electrocardiographic_type": datasets.Value("string"),
|
| 89 |
-
"maximum_heart_rate": datasets.Value("float32"),
|
| 90 |
-
"has_exercise_induced_angina": datasets.Value("bool"),
|
| 91 |
-
"depression_induced_by_exercise": datasets.Value("float32"),
|
| 92 |
-
"slope_of_peak_exercise": datasets.Value("float32"),
|
| 93 |
-
"number_of_major_vessels_colored_by_flourosopy": datasets.Value("int16"),
|
| 94 |
-
"thal": datasets.Value("float32"),
|
| 95 |
-
"has_hearth_disease": datasets.ClassLabel(num_classes=2, names=("no", "yes"))
|
| 96 |
-
},
|
| 97 |
"hungary": {
|
| 98 |
"age": datasets.Value("int8"),
|
| 99 |
"is_male": datasets.Value("bool"),
|
|
@@ -127,14 +76,8 @@ class HeartConfig(datasets.BuilderConfig):
|
|
| 127 |
|
| 128 |
class Heart(datasets.GeneratorBasedBuilder):
|
| 129 |
# dataset versions
|
| 130 |
-
DEFAULT_CONFIG = "
|
| 131 |
BUILDER_CONFIGS = [
|
| 132 |
-
HeartConfig(name="cleveland",
|
| 133 |
-
description="Heart for binary classification, dataset."),
|
| 134 |
-
HeartConfig(name="va",
|
| 135 |
-
description="Heart for binary classification, va dataset."),
|
| 136 |
-
HeartConfig(name="switzerland",
|
| 137 |
-
description="Heart for binary classification, switzerland dataset."),
|
| 138 |
HeartConfig(name="hungary",
|
| 139 |
description="Heart for binary classification, hungary dataset.")
|
| 140 |
]
|
|
@@ -158,12 +101,6 @@ class Heart(datasets.GeneratorBasedBuilder):
|
|
| 158 |
data.columns = _BASE_FEATURE_NAMES
|
| 159 |
data = self.preprocess(data, self.config.name)
|
| 160 |
|
| 161 |
-
print(data.head())
|
| 162 |
-
print(data.dtypes)
|
| 163 |
-
for f in data.columns:
|
| 164 |
-
print(f)
|
| 165 |
-
print(data[f].unique())
|
| 166 |
-
|
| 167 |
for row_id, row in data.iterrows():
|
| 168 |
data_row = dict(row)
|
| 169 |
|
|
@@ -176,47 +113,26 @@ class Heart(datasets.GeneratorBasedBuilder):
|
|
| 176 |
|
| 177 |
data[["age"]].applymap(int)
|
| 178 |
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
| 184 |
-
|
| 185 |
-
|
| 186 |
-
|
| 187 |
-
|
| 188 |
-
|
| 189 |
-
|
| 190 |
-
|
| 191 |
-
|
| 192 |
-
|
| 193 |
-
|
| 194 |
-
|
| 195 |
-
print(f)
|
| 196 |
-
data = data.astype({"is_male": bool, "has_exercise_induced_angina": bool,
|
| 197 |
-
"serum_cholesterol": float, "maximum_heart_rate": float,
|
| 198 |
-
"resting_blood_pressure": float, "fasting_blood_sugar": float})
|
| 199 |
-
else:
|
| 200 |
-
data = data[data.thal != "?"]
|
| 201 |
-
data = data[data.number_of_major_vessels_colored_by_flourosopy != "?"]
|
| 202 |
-
data = data[data.resting_blood_pressure != "?"]
|
| 203 |
-
data = data[data.fasting_blood_sugar != "?"]
|
| 204 |
-
data = data[data.rest_electrocardiographic_type != "?"]
|
| 205 |
-
data = data[data.maximum_heart_rate != "?"]
|
| 206 |
-
data = data[data.has_exercise_induced_angina != "?"]
|
| 207 |
-
|
| 208 |
-
for f in data.columns:
|
| 209 |
-
if "?" in data[f].values:
|
| 210 |
-
print(f)
|
| 211 |
-
data = data.astype({"is_male": bool, "has_exercise_induced_angina": bool,
|
| 212 |
-
"serum_cholesterol": float, "maximum_heart_rate": float,
|
| 213 |
-
"resting_blood_pressure": float, "fasting_blood_sugar": float,
|
| 214 |
-
"thal": float, "number_of_major_vessels_colored_by_flourosopy": float})
|
| 215 |
-
|
| 216 |
|
| 217 |
return data
|
| 218 |
|
| 219 |
def encode(self, feature, value):
|
| 220 |
if feature in _ENCODING_DICS:
|
| 221 |
return _ENCODING_DICS[feature][value]
|
| 222 |
-
raise ValueError(f"Unknown feature: {feature}")
|
|
|
|
| 40 |
|
| 41 |
# Dataset info
|
| 42 |
urls_per_split = {
|
|
|
|
| 43 |
"hungary": {"train": "https://huggingface.co/datasets/mstz/heart/raw/main/processed.hungarian.data"},
|
|
|
|
|
|
|
| 44 |
}
|
| 45 |
features_types_per_config = {
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 46 |
"hungary": {
|
| 47 |
"age": datasets.Value("int8"),
|
| 48 |
"is_male": datasets.Value("bool"),
|
|
|
|
| 76 |
|
| 77 |
class Heart(datasets.GeneratorBasedBuilder):
|
| 78 |
# dataset versions
|
| 79 |
+
DEFAULT_CONFIG = "hungary"
|
| 80 |
BUILDER_CONFIGS = [
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 81 |
HeartConfig(name="hungary",
|
| 82 |
description="Heart for binary classification, hungary dataset.")
|
| 83 |
]
|
|
|
|
| 101 |
data.columns = _BASE_FEATURE_NAMES
|
| 102 |
data = self.preprocess(data, self.config.name)
|
| 103 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 104 |
for row_id, row in data.iterrows():
|
| 105 |
data_row = dict(row)
|
| 106 |
|
|
|
|
| 113 |
|
| 114 |
data[["age"]].applymap(int)
|
| 115 |
|
| 116 |
+
data.drop("slope_of_peak_exercise", axis="columns", inplace=True)
|
| 117 |
+
data.drop("number_of_major_vessels_colored_by_flourosopy", axis="columns", inplace=True)
|
| 118 |
+
data.drop("thal", axis="columns", inplace=True)
|
| 119 |
+
data = data[data.serum_cholesterol != "?"]
|
| 120 |
+
|
| 121 |
+
data = data.infer_objects()
|
| 122 |
+
|
| 123 |
+
data = data[data.resting_blood_pressure != "?"]
|
| 124 |
+
data = data[data.fasting_blood_sugar != "?"]
|
| 125 |
+
data = data[data.rest_electrocardiographic_type != "?"]
|
| 126 |
+
data = data[data.maximum_heart_rate != "?"]
|
| 127 |
+
data = data[data.has_exercise_induced_angina != "?"]
|
| 128 |
+
|
| 129 |
+
data = data.astype({"is_male": bool, "has_exercise_induced_angina": bool,
|
| 130 |
+
"serum_cholesterol": float, "maximum_heart_rate": float,
|
| 131 |
+
"resting_blood_pressure": float, "fasting_blood_sugar": float})
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 132 |
|
| 133 |
return data
|
| 134 |
|
| 135 |
def encode(self, feature, value):
|
| 136 |
if feature in _ENCODING_DICS:
|
| 137 |
return _ENCODING_DICS[feature][value]
|
| 138 |
+
raise ValueError(f"Unknown feature: {feature}")
|