OneR算法_0(python数据挖掘入门与实践-实验2)

结果

代码

from sklearn.datasets import load_iris
import numpy as np
from collections import defaultdict
from operator import itemgetter
from sklearn.model_selection import train_test_split


#接收数据集,对应类别,特征索引,具体特征值
#返回特征索引具体特征值对分类最好的类别,以及分类错误的数量
def train_feature_value(X, y_true, feature_index, value):
    
    #统计 类别为feature_index且其特征值为value的个体
    class_counts=defaultdict(int)
    for sample,y in zip(X,y_true):
        if(sample[feature_index]==value):
            class_counts[y]+=1
    
    #排序
    sorted_class_counts=sorted(class_counts.items(),key=itemgetter(1),reverse=True)
    
    #符合要求个体最多的类别
    most_frequent_class=sorted_class_counts[0][0]
    
    #计算以该特征值分类符合要求个体最多的类别的错误率
    incorrect_predictions=[class_count for class_value,class_count in class_counts.items() if class_value!=most_frequent_class]
    error=sum(incorrect_predictions)
    
    
    return most_frequent_class,error

#接收数据集,对应类别,特征索引
#返回特征索引每一个特征值分类最好的类别,以及分类错误的数量
def train_on_feature(X, y_true, feature_index):
    values=set(X[:,feature_index])
    predictors={}
    errors=[]

    #计算某个特征其不同特征值分类能力
    for current_value in values:
        predictors[current_value],error=train_feature_value(X, y_true, feature_index, current_value)
        errors.append(error)
    
    return predictors,sum(errors)

#接收训练集
#返回预测器model
#model:feature为用于预测的特征的索引, predictor为特征的特征值对应的预测类别
#model={'feature': 2(特征索引), 'predictor': {0(特征值): 0(对应类别), 1(特征值): 2(类别)}}
def Training(Xd_train, Yd_train):
    all_predictors={}
    errors={}

    #计算每个特征的分类能力
    for feature_index in range(Xd_train.shape[1]):
        predictor,error=train_on_feature(Xd_train,Yd_train,feature_index)
        all_predictors[feature_index]=predictor
        errors[feature_index]=error

    best_feature,minimum_error=sorted(errors.items(),key=itemgetter(1),reverse=False)[0]
    model={'feature':best_feature,'predictor':all_predictors[best_feature]}

    return model

#接收测试集,预测模型
#返回预测集
def predict(Xd_test, model):
    feature_index=model['feature']
    predictor=model['predictor']
    y_predicted=np.array([predictor[int(sample[feature_index])] for sample in Xd_test])

    return y_predicted


#获取数据集
dataset=load_iris()
X=dataset.data
Y=dataset.target

#数据集离散化
attribute_means=X.mean(axis=0)
X_d=np.array(X>=attribute_means,dtype='int')

#数据集分割
Xd_train,X_test,Yd_train,Y_test=train_test_split(X_d,Y,random_state=14)

#预测模型训练
model=Training(Xd_train, Yd_train)
print("The predict model:{0}".format(model))

#预测模型评估
y_predicted=predict(X_test, model)
accuracy=np.mean(y_predicted==Y_test)*100
print("The test accuracy is {:.1f}%".format(accuracy))