MATLAB Answers

Matlab Clustering technique with textual data

5 views (last 30 days)
Brian
Brian on 12 Dec 2016
Answered: mizuki on 20 Dec 2016
Hi, I am trying to figure out the best way to cluster numeric information (stock returns) using a series of textual information. For instance, let's say I have 10 sectors with of stock returns that I'd like to cluster to 3 distinct groups. My first thought was to use the K-means clustering algorithm from the "Stats and ML" toolbox however, it doesn't take textual information as a descriptor.
Please advise.
Example data set
Industry, Return
Financials,2%
Consumer Disc,3%
Consumer Staples,4.5%
Energy,1%
Health Care,1.5%
Industrials,2.2%
Info Tech,3.7%
Materials,4.8%
Telecom,-2%
Utilities,-1%
  1 Comment
Brian
Brian on 16 Dec 2016
Any ideas on this from statistical experts?

Sign in to comment.

Answers (1)

mizuki
mizuki on 20 Dec 2016
Make the textual data categorical to reduce information.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by