nav[aria-label="Primary Navigation"] { padding: 0; & ul { list-style: none; width: 100%; display: flex; flex-direction: row; justify-content: start; align-items: start; gap: 30px; padding: 0; & li { margin: 0; } & ul li { list-style: none; } } }

Siemens Community Catalyst Program

The Siemens Community Catalyst program was co-created with our community to acknowledge technology leaders who consistently contribute to the Siemens Community. Nominations are accepted on a rolling basis.

Nominate Now

"Bigram Document Vector"

D_M

Hi,

I want to create document vector consisting only of bigrams.

For this I am first saving the wordlist using the following operators:-

TextInput
StringTokenizer

and then I am using

TextInput
StringTokenizer
TermNgramGenerator
StopWordFilterFile (using the previously saved wordlist.)

Is there any better way of doing this?

Find more posts tagged with

AI Studio

Text Mining + NLP

Accepted answers

All comments

D_M

land

Hi,
I guess it is. Why don't you just use it this way:
TextInput
StringTokenizer
TermNgramGenerator

The resulting vector will only contain the bi-grams, since it builds the vector from the tokens generated by all inner operators. If no token of a complete word is contained, it will not be part of the vector.
Or did I misunderstand you completely?

Greetings,
Sebastian

D_M

Thanks Sebastian for replying.

I tried using only

TextInput
StringTokenizer
TermNGramGenerator

The problem I am facing is that along with the bigrams, unigrams are also coming to the document vector. I want only bigrams not unigrams. So to prevent this I have to use the StopWordFilter to remove the unigrams.

Plez let me know if I can achieve this in a much better way?

haddock

What Ho D.M !

Not really understanding much about anything I looked up on Wikipedia to understand what a bigram was, and found the following...

An n-gram is a subsequence of n items from a given sequence. The items in question can be phonemes, syllables, letters, words or base pairs according to the application.

Here http://en.wikipedia.org/wiki/N-gram

What is the context of your application?

D_M

Sorry, if my question is not clear.

For me the bigram should composed of sequence of words.

e.g. For - "the dog smelled like a skunk" bigrams should be xx_the, the_dog, dog_smelled, smelled_like, like_a, a_shrunk, shrunk_xx.

land

Hi,
with RapidMiner 5.0 the result should exactly look like what you are expecting it to be. Otherwise there's no possibility to change this, but you could filter the not desired results using the example filter.
You might specify a regular expression for filtering the attribute according to their names.

Greetings,
Sebastian

D_M

thanks Sebastian for replying.