Clustering and Distance/Dissimilarity Matrrix based on String/Integer sequences in Python

Question

I have customer's data based on his stay in the shop. The shop has 4 zones; zone 1,2,3 and 4. Now every 2 minutes, I get his reading as 10 numbers based on which zone he is in. EX:

1-1-1-1-1-1-1-1-3-3-2
4-4-3-3-3-3-3-2-1-3-3
3-4-1-2-2-3-1-4-2-1-4

Basically, I expect that there are customers who mostly are in a particular zone and they are clustered accordingly. So, in the first sequence, the customer seems to prefer zone 1, the next one zone 3 and the last one is like noise.

All I am feeding to the program is a bunch of sequences (unlabeled). How do I generate a distance/dissimilarity matrix that calculates the distances between each sequence in Python?

score 0 · Accepted Answer · answered Sep 09 '19 at 03:43

0

After a little bit of digging, I came across the textdistance library in python.

https://pypi.org/project/textdistance/

It seems to be working well for this problem, even though my input is a sequence of integers.

answered Sep 09 '19 at 03:43

Sakib Shahriar

121
1
12

which distance method did you use in the end? – Shihab Shahriar Khan Nov 06 '19 at 14:00
I have the Ratcliff-Obershelp similarity and got reasonably good results – Sakib Shahriar Nov 07 '19 at 14:39

score 0 · Answer 2 · answered Sep 09 '19 at 03:55

You can use cosine or euclidean distances to calculate the distance.

https://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.spatial.distance.cosine.html

https://scikit-learn.org/stable/modules/generated/sklearn.metrics.pairwise.euclidean_distances.html

Clustering and Distance/Dissimilarity Matrrix based on String/Integer sequences in Python

2 Answers2