The oneR algorithm returns a rule list that splits on only one (usually continuous) feature It works by building a greedy rule list using only one feature at a time, and then returning the rule list with the highest accuracy
Expand source code
'''The oneR algorithm returns a rule list that splits on only one (usually continuous) feature
It works by building a greedy rule list using only one feature at a time, and then returning
the rule list with the highest accuracy
'''
import math
import numpy as np
from copy import deepcopy
from sklearn.base import BaseEstimator
from sklearn.base import BaseEstimator, ClassifierMixin, RegressorMixin
from sklearn.utils.multiclass import check_classification_targets, unique_labels
from sklearn.utils.validation import check_X_y, check_array, check_is_fitted
from imodels import GreedyRuleListClassifier
from imodels.rule_list.rule_list import RuleList
from imodels.util.arguments import check_fit_arguments
class OneRClassifier(GreedyRuleListClassifier):
def __init__(self, max_depth=5, class_weight=None, criterion='gini'):
self.max_depth = max_depth
self.feature_names_ = None
self.class_weight = class_weight
self.criterion = criterion
self._estimator_type = 'classifier'
def fit(self, X, y, feature_names=None):
"""Fit oneR
"""
X, y, feature_names = check_fit_arguments(self, X, y, feature_names)
ms = []
accs = np.zeros(X.shape[1])
for col_idx in range(X.shape[1]):
x = X[:, col_idx].reshape(-1, 1)
m = GreedyRuleListClassifier(max_depth=self.max_depth, class_weight=self.class_weight,
criterion=self.criterion)
feat_names_single = [self.feature_names_[col_idx]]
m.fit(x, y, feature_names=feat_names_single)
accs[col_idx] = np.mean(m.predict(x) == y)
ms.append(m)
# print('acc', feat_names_single[0], f'{accs[col_idx]:0.2f}')
col_idx_best = np.argmax(accs)
self.rules_ = ms[col_idx_best].rules_
self.complexity_ = len(self.rules_)
# need to adjust index_col since was fitted with only 1 col
for rule in self.rules_:
if 'index_col' in rule:
rule['index_col'] += col_idx_best
self.depth = len(self.rules_)
return self
Classes
class OneRClassifier (max_depth=5, class_weight=None, criterion='gini')
-
Base class for all estimators in scikit-learn.
Notes
All estimators should specify all the parameters that can be set at the class level in their
__init__
as explicit keyword arguments (no*args
or**kwargs
).Params
max_depth Maximum depth the list can achieve criterion: str Criterion used to split 'gini', 'entropy', or 'log_loss'
Expand source code
class OneRClassifier(GreedyRuleListClassifier): def __init__(self, max_depth=5, class_weight=None, criterion='gini'): self.max_depth = max_depth self.feature_names_ = None self.class_weight = class_weight self.criterion = criterion self._estimator_type = 'classifier' def fit(self, X, y, feature_names=None): """Fit oneR """ X, y, feature_names = check_fit_arguments(self, X, y, feature_names) ms = [] accs = np.zeros(X.shape[1]) for col_idx in range(X.shape[1]): x = X[:, col_idx].reshape(-1, 1) m = GreedyRuleListClassifier(max_depth=self.max_depth, class_weight=self.class_weight, criterion=self.criterion) feat_names_single = [self.feature_names_[col_idx]] m.fit(x, y, feature_names=feat_names_single) accs[col_idx] = np.mean(m.predict(x) == y) ms.append(m) # print('acc', feat_names_single[0], f'{accs[col_idx]:0.2f}') col_idx_best = np.argmax(accs) self.rules_ = ms[col_idx_best].rules_ self.complexity_ = len(self.rules_) # need to adjust index_col since was fitted with only 1 col for rule in self.rules_: if 'index_col' in rule: rule['index_col'] += col_idx_best self.depth = len(self.rules_) return self
Ancestors
- GreedyRuleListClassifier
- sklearn.base.BaseEstimator
- RuleList
- sklearn.base.ClassifierMixin
Methods
def fit(self, X, y, feature_names=None)
-
Fit oneR
Expand source code
def fit(self, X, y, feature_names=None): """Fit oneR """ X, y, feature_names = check_fit_arguments(self, X, y, feature_names) ms = [] accs = np.zeros(X.shape[1]) for col_idx in range(X.shape[1]): x = X[:, col_idx].reshape(-1, 1) m = GreedyRuleListClassifier(max_depth=self.max_depth, class_weight=self.class_weight, criterion=self.criterion) feat_names_single = [self.feature_names_[col_idx]] m.fit(x, y, feature_names=feat_names_single) accs[col_idx] = np.mean(m.predict(x) == y) ms.append(m) # print('acc', feat_names_single[0], f'{accs[col_idx]:0.2f}') col_idx_best = np.argmax(accs) self.rules_ = ms[col_idx_best].rules_ self.complexity_ = len(self.rules_) # need to adjust index_col since was fitted with only 1 col for rule in self.rules_: if 'index_col' in rule: rule['index_col'] += col_idx_best self.depth = len(self.rules_) return self