使用orange3-associate做分类关联规则

最新推荐文章于 2025-03-28 14:14:12 发布

njzhuming

最新推荐文章于 2025-03-28 14:14:12 发布

阅读量1.2k

点赞数

分类专栏： python 文章标签： python 数据挖掘

本文链接：https://blog.csdn.net/njzhuming/article/details/113856315

版权

python 专栏收录该内容

9 篇文章

订阅专栏

该博客介绍了如何使用Orange库进行关联规则分析。通过读取'zoo'数据集，设置40%的支持度阈值，提取频繁项集，并基于这些项集生成规则。最终得到了一系列关于动物特征与分类（哺乳动物）的关联规则，如具备特定特征的动物群体100%属于哺乳动物等。这些规则展示了特征间的关联性。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

分类关联规则示例：

读入数据

import Orange

data = Orange.data.Table('zoo')
data

转换

X, mapping = OneHot.encode(data, include_class=True)
itemsets = dict(frequent_itemsets(X, .4))#40%支持度
len(itemsets)
class_items = {item
               for item, var, _ in OneHot.decode(mapping, data, mapping)
               if var is data.domain.class_var}
sorted(class_items)
data.domain.class_var.values#查看分类

产生规则

rules = [(P, Q, supp, conf)
          for P, Q, supp, conf in association_rules(itemsets, .8)
          if len(Q) == 1 and Q & class_items]
len(rules)

得到如下规则：

[(frozenset({2, 7, 17, 19, 20}), frozenset({41}), 41, 1.0),
 (frozenset({2, 7, 17, 19}), frozenset({41}), 41, 1.0),
 (frozenset({2, 7, 17, 20}), frozenset({41}), 41, 1.0),
 (frozenset({2, 7, 19, 20}), frozenset({41}), 41, 1.0),
 (frozenset({2, 17, 19, 20}), frozenset({41}), 41, 0.8723404255319149),
 (frozenset({7, 17, 19, 20}), frozenset({41}), 41, 1.0),
 (frozenset({2, 7, 17}), frozenset({41}), 41, 1.0),
 (frozenset({2, 7, 19}), frozenset({41}), 41, 1.0),
 (frozenset({2, 17, 19}), frozenset({41}), 41, 0.8367346938775511),
 (frozenset({7, 17, 19}), frozenset({41}), 41, 1.0),
 (frozenset({2, 7, 20}), frozenset({41}), 41, 1.0),
 (frozenset({7, 17, 20}), frozenset({41}), 41, 1.0),
 (frozenset({7, 19, 20}), frozenset({41}), 41, 1.0),
 (frozenset({2, 7}), frozenset({41}), 41, 1.0),
 (frozenset({7, 17}), frozenset({41}), 41, 1.0),
 (frozenset({7, 19}), frozenset({41}), 41, 1.0),
 (frozenset({7, 20}), frozenset({41}), 41, 1.0),
 (frozenset({7}), frozenset({41}), 41, 1.0)]

查看详细内容

names = {item: '{}={}'.format(var.name, val)
          for item, var, val in OneHot.decode(mapping, data, mapping)}
for ante, cons, supp, conf in rules[:5]:
    print(', '.join(names[i] for i in ante), '-->',
           names[next(iter(cons))],
           '(supp: {}, conf: {})'.format(supp, conf))

结果如下：

feathers=0, milk=1, backbone=1, breathes=1, venomous=0 --> type=mammal (supp: 41, conf: 1.0)
backbone=1, feathers=0, breathes=1, milk=1 --> type=mammal (supp: 41, conf: 1.0)
backbone=1, feathers=0, venomous=0, milk=1 --> type=mammal (supp: 41, conf: 1.0)
feathers=0, breathes=1, venomous=0, milk=1 --> type=mammal (supp: 41, conf: 1.0)
backbone=1, feathers=0, breathes=1, venomous=0 --> type=mammal (supp: 41, conf: 0.8723404255319149)