I am currently trying to develop a classifier in python using Naive Bayes technique. I need a dataset so that I can train it. My classifier would classify a new document given to it into one of the four categories : Science and technology, Sports, politics, Entertainment. Can anybody please help me find a dataset for this. I've been stuck on this problem for quite some time now. Any help would be greatly appreciated.
2
-
$\begingroup$ What type of documents are you intending to look at? Length wise are you looking at Tweets, Tumblr posts, articles, etc? $\endgroup$– Aonghus CollinsApr 24, 2015 at 15:43
-
$\begingroup$ I am Looking for articles preferably. $\endgroup$– user3848891Apr 24, 2015 at 19:03
Add a comment
|
3 Answers
$\begingroup$
$\endgroup$
You have at least three options:
Use some of many available datasets (for example: BBC documents); if you need more, simply go to Google Scholar or any similar service and search for classification news politics sports. In the articles you may find many references to available datasets;
Crawl any news service, and use clustering techniques to group articles into clusters (this often separates the articles along their domain, such as politics, sports etc.), and label the articles by their affiliation to a cluster;
Crawl politics-specific, sports-specific news services and use them as labeled datasets.
$\begingroup$
$\endgroup$
The UCI Machine Learning repo is my go to place to find data sets to work on across a really broad range of topics. The data sets are already cleaned and tagged with the task they were originally purposed for. If you google the data set you'll often find papers citing them or forums with questions that you may have yourself. There's a number of text based classification tasked data sets in there.
