1
$\begingroup$

This question is related to text analytics. I have text files which contain customer feedback for a chain of retail stores. My objective is to extract 5 main topics or entities from that data and score each store based on those entities and customer feedback corresponding to that store. Can someone tell me how i can do this?

Pointing me towards some reading material also would be very helpful!!

[I am a beginner in text analytics!!]

$\endgroup$
0
$\begingroup$

Kan Nishida did a few blog posts that might help you start, especially for the preparation of data and the clustering part :

  1. Demystifying Text Analytics part 1—  Preparing Document and Term Data for Text Mining in R,
  2. Demystifying Text Analytics part 2 —  Quantifying Documents by Calculating TF-IDF in R,
  3. Demystifying Text Analytics part 3—  Finding Similar Documents with Cosine Similarity in R,
  4. Demystifying Text Analytics part 4— Dimensionality Reduction and Clustering in R,
  5. Demystifying Text Analytics part 5— Finding the most relevant terms for each cluster.

There are data preparation techniques that are specific to text mining. It's very important to not underestimate this step as it will be crucial to extract relevant clusters.

$\endgroup$
0
$\begingroup$

Look at Latent Dirichlet Allocation (LDA). This would give you topics and a probability for each which you could use to derive a score.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.