Datasets:
Upload compas.py
Browse files
compas.py
CHANGED
|
@@ -73,17 +73,15 @@ _BASE_FEATURE_NAMES = [
|
|
| 73 |
"decile_score",
|
| 74 |
"number_of_juvenile_misdemeanors",
|
| 75 |
"number_of_other_juvenile_offenses",
|
| 76 |
-
"
|
| 77 |
"days_before_screening_arrest",
|
| 78 |
"is_recidivous",
|
|
|
|
| 79 |
"days_of_recidividity_after_arrest",
|
| 80 |
-
"days_in_jail_before_recidividity",
|
| 81 |
-
"days_in_custody",
|
| 82 |
"is_violent_recidivous",
|
| 83 |
"violence_decile_score",
|
| 84 |
-
"
|
| 85 |
-
"
|
| 86 |
-
"two_year_recidivous"
|
| 87 |
]
|
| 88 |
DESCRIPTION = "COMPAS dataset for recidivism prediction."
|
| 89 |
_HOMEPAGE = "https://github.com/propublica/compas-analysis"
|
|
@@ -262,6 +260,8 @@ class Compas(datasets.GeneratorBasedBuilder):
|
|
| 262 |
data.drop("compas_screening_date", axis="columns", inplace=True)
|
| 263 |
data.drop("decile_score.1", axis="columns", inplace=True)
|
| 264 |
data.drop("priors_count.1", axis="columns", inplace=True)
|
|
|
|
|
|
|
| 265 |
|
| 266 |
# drop nan values
|
| 267 |
data = data[~data.days_b_screening_arrest.isna()]
|
|
|
|
| 73 |
"decile_score",
|
| 74 |
"number_of_juvenile_misdemeanors",
|
| 75 |
"number_of_other_juvenile_offenses",
|
| 76 |
+
"number_of_prior_offenses",
|
| 77 |
"days_before_screening_arrest",
|
| 78 |
"is_recidivous",
|
| 79 |
+
# "days_in_jail_before_recidividity",
|
| 80 |
"days_of_recidividity_after_arrest",
|
|
|
|
|
|
|
| 81 |
"is_violent_recidivous",
|
| 82 |
"violence_decile_score",
|
| 83 |
+
"two_year_recidivous",
|
| 84 |
+
"days_in_custody",
|
|
|
|
| 85 |
]
|
| 86 |
DESCRIPTION = "COMPAS dataset for recidivism prediction."
|
| 87 |
_HOMEPAGE = "https://github.com/propublica/compas-analysis"
|
|
|
|
| 260 |
data.drop("compas_screening_date", axis="columns", inplace=True)
|
| 261 |
data.drop("decile_score.1", axis="columns", inplace=True)
|
| 262 |
data.drop("priors_count.1", axis="columns", inplace=True)
|
| 263 |
+
data.drop("c_case_number", axis="columns", inplace=True)
|
| 264 |
+
data.drop("c_days_from_compas", axis="columns", inplace=True)
|
| 265 |
|
| 266 |
# drop nan values
|
| 267 |
data = data[~data.days_b_screening_arrest.isna()]
|