如何计算两个字符串向量之间的余弦相似度

Question

I have 2 vectors of dimensions 6 and I would like to have a number between 0 and 1.

a=c("HDa","2Pb","2","BxU","BuQ","Bve")

b=c("HCK","2Pb","2","09","F","G")

Can anyone explain what I should do?

Answer 1

using the lsa package and the manual for this package

# create some files
library('lsa')
td = tempfile()
dir.create(td)
write( c("HDa","2Pb","2","BxU","BuQ","Bve"), file=paste(td, "D1", sep="/"))
write( c("HCK","2Pb","2","09","F","G"), file=paste(td, "D2", sep="/"))

# read files into a document-term matrix
myMatrix = textmatrix(td, minWordLength=1)

EDIT: show how is the mymatrix object

myMatrix
#myMatrix
#       docs
#  terms D1 D2
#    2    1  1
#    2pb  1  1
#    buq  1  0
#    bve  1  0
#    bxu  1  0
#    hda  1  0
#    09   0  1
#    f    0  1
#    g    0  1
#    hck  0  1

# Calculate cosine similarity
res <- lsa::cosine(myMatrix[,1], myMatrix[,2])
res
#0.3333

Answer 2

You need a dictionary of possible terms first and then convert your vectors to binary vectors with a 1 in the positions of the corresponding terms and 0 elsewhere. If you name the new vectors a2 and b2 , you can calculate the cosine similarly with cor(a2, b2) , but notice the cosine similarly is between -1 and 1. You could map it to [0,1] with something like this: 0.5*cor(a2, b2) + 0.5

Answer 3

CSString_vector <- c("Hi Hello","Hello");
corp <- tm::VCorpus(VectorSource(CSString_vector));
controlForMatrix <- list(removePunctuation = TRUE,wordLengths = c(1, Inf), weighting = weightTf)
dtm <- DocumentTermMatrix(corp,control = controlForMatrix);
matrix_of_vector = as.matrix(dtm);
res <- lsa::cosine(matrix_of_vector[1,], matrix_of_vector[2,]);

could be the better one for the larger data set.

Answer 4

Advanced form of embedding might help you to get better output. Please check the following code. It is a Universal sentence encode model that generates the sentence embedding using transformer-based architecture.

from absl import logging
import tensorflow as tf
import tensorflow_hub as hub
import matplotlib.pyplot as plt
import numpy as np
import os
import pandas as pd
import re
import seaborn as sns

module_url = "https://tfhub.dev/google/universal-sentence-encoder/4"
model = hub.load(module_url)
print ("module %s loaded" % module_url)
def embed(input):
  return model([input])

paragraph = [
    "Universal Sentence Encoder embeddings also support short paragraphs. ",
    "Universal Sentence Encoder support paragraphs"]
messages = [paragraph]

print(np.inner( embed(paragraph[0]) , embed(paragraph[1])))