Clustering sequences of data with different lengths using ngrams.

I am trying to cluster sequences of data that have different lengths. 

I am treating the element of my sequence as a word. In other words, a list is a sentence.

To be able to achive this I am thinking of extracting ngrams like my code below.

However, I keep getting following error message:

“Arguments to IMPORT must either end with ".*" or else specify a fully qualified class name: "nltk" fails this test.”

1. How can I solve this error?

2. Am I going toward right direction?

I am a novice Matlab user and I hope I am not asking too basic questions.

Following is my Matlab code:

-------------------------------------------------------

A = [1.2,1.5; 2.4,0.5; 3.2,1.5];

B = [1.0,1.0; 2.0,0; 3.0,1.0; 4.0,0.5];

C = [1.1,1.2; 2.2,1.3; 3.3,1.5];

D = [1.3,1.3; 2.1,1.4; 3.2,1.7; 4.3, 2.0];

E = [1.3,1.8; 2.3,0.4; 3.1,1.6; 4.3,0.8];

figure()

plot(A(:,1), A(:,2)) % first column is the time-values

hold on;

plot(B(:,1), B(:,2))

hold on;

plot(E(:,1), E(:,2))

figure()

plot(C(:,1), C(:,2)) % first column is the time-values

hold on;

plot(D(:,1), D(:,2))

import nltk

from nltk import ngrams

aa = A(:)'

bb = B(:)'

cc = C(:)'

dd = D(:)'

ee = E(:)'

bb = list()

bb append(str(','.join(str(e) for e in ['x' + str(e) for e in aa])))

bb append(str(','.join(str(e) for e in ['x' + str(e) for e in bb])))

bb append(str(','.join(str(e) for e in ['x' + str(e) for e in cc])))

bb append(str(','.join(str(e) for e in ['x' + str(e) for e in dd])))

bb append(str(','.join(str(e) for e in ['x' + str(e) for e in ee])))

from sklearn.feature_extraction.text import CountVectorizer

vectorizer = CountVectorizer()

X = vectorizer.fit_transform(bb)

X.toarray()

Answer 1

[Prabhanjan Mentla]

NLTK is not supported in MATLAB however one can run python functions in MATLAB by following this. There is also text analytics toolbox in MATLAB which deals text data.

https://www.mathworks.com/products/text-analytics.html

[SOPAE YI]

Thank you so much for crrecting me, sir.

My data is not text.

I am working on spoken sentence (or an utterance) especialy, pitch of a sentence.

Can I use NLTK to cluster sequences of (pitch) data that have different lengths?