2
$\begingroup$

I'm building a system able to classify emails into different categories (positive, negative, out of office, etc...) and I'm looking for a dataset of already classified emails to avoid hand classification on a 70k database.

I know it exists the Enron email dataset but do you know if it exists a version of this dataset with classified emails ? Or any other already classified email dataset ?

Improve this question
$\endgroup$
1
  • 1
    $\begingroup$ I assume this is a question for the Open Data SE. $\endgroup$ Commented Nov 15, 2016 at 17:03

2 Answers 2

Reset to default
2
+50
$\begingroup$

You can download the corpus from this site. To the best of my knowledge this is the most complete email corpus available. A project to label a subset of this email corpus can be found on this UC Berkley site. I am not sure though whether these emails have the right training labels for you.

Improve this answer
$\endgroup$
1
  • $\begingroup$ Even if categories aren't exactly what I'm looking for It's a really good working basis! I'll just wait one or two days before accepting your answer in case a better choice is given to me. $\endgroup$ Commented Nov 18, 2016 at 9:13
0
$\begingroup$

The Enron Corpus: A New Dataset for Email Classification Research paper describes the kind of data set you want.

The paper mentions the following link to download the data set:

https://www.cs.cmu.edu/~./enron/

Additionally, the paper also mentions various other papers which have used smaller data sets related to email classification which may not be of much use given this larger dataset.

Improve this answer
$\endgroup$
3
  • $\begingroup$ No. This paper is about the number of mail per user, the number of message per thread and the number of folder per user but not about mail content classification. $\endgroup$ Commented Nov 15, 2016 at 15:58
  • $\begingroup$ @JérémyPouyet The paper does mention the task of classifying emails into folders. Please go through section 2 on page 3. $\endgroup$ Commented Nov 15, 2016 at 16:18
  • $\begingroup$ @JérémyPouyet I have added the link to the dataset from the paper. $\endgroup$ Commented Nov 15, 2016 at 16:35

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.