Sign Up to our social questions and Answers Engine to ask questions, answer people’s questions, and connect with other people.
Login to our social questions & Answers Engine to ask questions answer people’s questions & connect with other people.
Lost your password? Please enter your email address. You will receive a link and will create a new password via email.
Please briefly explain why you feel this question should be reported.
Please briefly explain why you feel this answer should be reported.
Please briefly explain why you feel this user should be reported.
I am using libsvm to predict sentiment. I wanted to know what format the input has to be in
assuming I am using word count.
[label] [index]:[value] [index]:[value]
That is required format from libsvm. So does that mean I just have two labels ( one for positive and one for negative), the index would be each word under that label and the value would be the frequency of each word ?
Does this also mean I need to store the mapping of word to index to use in my test set ?
LIBSVM uses the so called “sparse” format where zero values do not need to be stored. Hence a data with attributes
5 0 2 0
is represented as
1:5 3:2
Therefore, you only need to specifiy the index and the value of nonzero attributes.
Labels stand in the first column. For binary cases you may use +1 for positive and -1 for negative samples. By the way, you are not limited to only 2 labels. You can use other numbers (e.g. 1,2,3,4,5,…)