Grouping similar news contents together like in GOOGLE NEWS

Viewed 2457

I am unable to manage the RSS feeds easily due to an overwhelming number of new stories / similar news contents posted in various news sites. For subjects such as world news and business news, many of the stories are redundant, adding a burden to readers to sort out which stories they've already read. To deal with the twin problems of flooding and redundancy, i need to develop an code that reduces the number of items to read and uses the overlapping information to divine interesting topics.

it would be easier if i am able to Grouping similar news contents together like in GOOGLE NEWS / StackOverflow and present it to the users.

6 Answers

I have has success doing this using by dumping all the articles into Elsasticsearch and doing a more_like_this query. This works surprisingly well. It just took some fine-tuning to get some of the settings right. You can also use a free Elasticsearch hosted instance at bonsai.io

Related