01_字符串处理-----05_相似性度量

# NLTK中的nltk.metrics包用于提供各种评估或相似性度量
from __future__ import print_function
from nltk.metrics import *


def main1():
    training = 'PERSON OTHER PERSON OTHER OTHER ORGANIZATION'.split()
    testing = 'PERSON OTHER OTHER OTHER OTHER OTHER'.split()
    print(training)
    print(testing)
    print("ACCURACY:")
    print(accuracy(training, testing))
    print("===" * 25)
    trainset = set(training)
    testset = set(testing)
    print(trainset)
    print(testset)
    print("PRECISION:")
    print(precision(trainset, testset))
    print("RECALL:")
    print(recall(trainset, testset))
    print("F_MEASURE:")
    print(f_measure(trainset, testset))

main1()

# 执行结果
['PERSON', 'OTHER', 'PERSON', 'OTHER', 'OTHER', 'ORGANIZATION']
['PERSON', 'OTHER', 'OTHER', 'OTHER', 'OTHER', 'OTHER']
ACCURACY:
0.6666666666666666
===========================================================================
{'ORGANIZATION', 'PERSON', 'OTHER'}
{'PERSON', 'OTHER'}
PRECISION:
1.0
RECALL:
0.6666666666666666
F_MEASURE:
0.8

1.5.1 使用编辑距离算法执行相似性度量

import nltk
from nltk.metrics import *

def main2():
    # 3步
    print(edit_distance("relate", "relation"))
    # 7步
    print(edit_distance("suggestion", "calculation"))

main2()

# 执行结果
3
7

1.5.2 使用Jaccard系数执行相似性度量

# 两个集合 X 和 Y 交集的相似程度
import nltk
from nltk.metrics import *

def main3():
    X1 = {20, 30}
    Y2 = {20, 30}
    print("完全相同：")
    print(jaccard_distance(X1, Y2))

    X1 = {20, 30}
    Y2 = {10, 40}
    print("完全不同")
    print(jaccard_distance(X1, Y2))

    X1 = {20, 30}
    Y2 = {20, 40}
    print("部分相同")
    print(jaccard_distance(X1, Y2))

main3()

# 执行结果
完全相同：
0.0
完全不同
1.0
部分相同
0.6666666666666666

1.5.3 使用Smith Waterman 距离算法执行相似度计算

# 类似于编辑距离算法
# 在nltk.metrics包中没有找到相关模块

1.5.4 其他字符串相似性度量

# 二进制距离
# 两个标签相同，返回值为0.0， 否则，它的返回值为1.0
import nltk
from nltk.metrics import *

def main4():
    X = {30, 50}
    Y = {30, 50}
    print("相同")
    print(binary_distance(X, Y))
    X = {10, 20, 30, 40}
    Y = {30, 50, 70}
    print("不同")
    print(binary_distance(X, Y))

main4()

# 执行结果
相同
0.0
不同
1.0

# 多个标签,Masi距离基于部分协议
# 有详细算法
import nltk
from nltk.metrics import *

def main5():
    X = {30, 40}
    Y = {30, 40, 50, 60}
    print(masi_distance(X, Y))

main5()

# 执行结果
0.335