CODEWITHSUNDEEP
rcosine-similarity
Ozgur Alptekın
Ozgur Alptekın

Reputation: 505

How to calculate cosine similarity between vector and each rows of data frame in R?

this is my dataframe a:

                    ui 194635691 194153563 177382028 177382031 195129144 196972549 196258704 194907960 196950156 194139014 153444738
1 56320e0e55e89c3e14e26d3d      0.00      0.00      0.00      0.00      0.00      0.00      0.00      0.01     0.000         0         0
2 563734c3b65dd40e340eaa56      0.01      0.00      0.00      0.00      0.00      0.00      0.00      0.00     0.000         0         0
3 563e12657d4c410c5832579c      0.00      0.00      0.01      0.01      0.00      0.00      0.00      0.00     0.000         0         0
4 565181854c24b410e4891e11      0.00      0.00      0.00      0.00      0.00      0.01      0.00      0.00     0.000         0         0
5 5651b53fec231f1df8482d23      0.00      0.00      0.00      0.00      0.00      0.00      0.00      0.00     0.027         0         0
6 56548df4b84c321fe4cdfb8f      0.00      0.01      0.00      0.00      0.00      0.00      0.00      0.00     0.000         0         0
7 56549946735e782a885957e6      0.00      0.00      0.00      0.00      0.08      0.00      0.00      0.00     0.000         0         0
8 56549f9bb84c321fe4ce7a37      0.00      0.01      0.00      0.00      0.00      0.00      0.00      0.00     0.000         0         0
9 5654a35a735e782a8859a053      0.00      0.00      0.00      0.00      0.00      0.00      0.01      0.00     0.000         0         0

What I want to do here is calculate the cosine similarity between useridvector and each row of data frame a but without first column(ui) I have tried following code so far:

  user_id=actions_slippers$ui[i]#user_id is coming from another dataframe called action_slippers
  useridvector=a[a$ui %in% user_id, ]
  p=as.vector(cosine(t(a[,2:ncol(a)]))[,1])# this measures cosine similarity between first row of dataframe a and each other of rows from dataframe a

but I want to calculate cosine similarity between useridvector and each row of dataframe a without first column. useridvector looks like this:

 ui 194635691 194153563 177382028 177382031 195129144 196972549 196258704 194907960 196950156 194139014 153444738
 5651b53fec231f1df8482d23         0         0         0         0         0        0         0         0     0.027         0         0

Can anyone tell me how to do this?

Upvotes: 1

Views: 3632

Answers (1)

Dylan
Dylan

Reputation: 1365

cosine{lsa} works. I'd like to share my try.

suppose you save the data in a dataframe like:

> data
                        ui X194635691 X194153563 X177382028 X177382031 X195129144 X196972549 X196258704 X194907960 X196950156 X194139014 X153444738
1 56320e0e55e89c3e14e26d3d       0.00       0.00       0.00       0.00       0.00       0.00       0.00       0.01      0.000          0          0
2 563734c3b65dd40e340eaa56       0.01       0.00       0.00       0.00       0.00       0.00       0.00       0.00      0.000          0          0
3 563e12657d4c410c5832579c       0.00       0.00       0.01       0.01       0.00       0.00       0.00       0.00      0.000          0          0
4 565181854c24b410e4891e11       0.00       0.00       0.00       0.00       0.00       0.01       0.00       0.00      0.000          0          0
5 5651b53fec231f1df8482d23       0.00       0.00       0.00       0.00       0.00       0.00       0.00       0.00      0.027          0          0
6 56548df4b84c321fe4cdfb8f       0.00       0.01       0.00       0.00       0.00       0.00       0.00       0.00      0.000          0          0
7 56549946735e782a885957e6       0.00       0.00       0.00       0.00       0.08       0.00       0.00       0.00      0.000          0          0
8 56549f9bb84c321fe4ce7a37       0.00       0.01       0.00       0.00       0.00       0.00       0.00       0.00      0.000          0          0
9 5654a35a735e782a8859a053       0.00       0.00       0.00       0.00       0.00       0.00       0.01       0.00      0.000          0          0

Using data[,-1] or subset.data.frame(data, select = names(data)[-1] to eliminate the first column,then convert to matrix and use the cosine{lsa}

> res <- lsa::cosine(t(as.matrix(data[, -1])))
> res
      [,1] [,2] [,3] [,4] [,5] [,6] [,7] [,8] [,9]
 [1,]    1    0    0    0    0    0    0    0    0
 [2,]    0    1    0    0    0    0    0    0    0
 [3,]    0    0    1    0    0    0    0    0    0
 [4,]    0    0    0    1    0    0    0    0    0
 [5,]    0    0    0    0    1    0    0    0    0
 [6,]    0    0    0    0    0    1    0    1    0
 [7,]    0    0    0    0    0    0    1    0    0
 [8,]    0    0    0    0    0    1    0    1    0
 [9,]    0    0    0    0    0    0    0    0    1

PS: install the lsa package and see ?cosine for detail info

============================ update =====

The resulting matrix is like, 

      user1 **user2** user3 **user4**
user1   1    0 
user2        1
user3    ...       1
user4

where element(i,j) means the similarity between user i and user j. and if your userid has 2 users say user 2 and user 4. Then you want to find the similarity between these 2 users to other users. which is a sub matrix of the entire similarity matrix. Then use res[, c(2,4)] to obtain the desired matrix.

Upvotes: 5

Related Questions