With 1.7 million freelancers, we'll match you with the perfect talent.
The client has made the following changes to the job.
Client prefers freelancers from:
You are still able to submit a proposal for this job.
The client prefers freelancers from
a different location.
Using Python, NLP, web crawling, and any other free tools...
The attached spreadsheet has a list of image urls, and for each image url, approximately 240 "tags". These tags were adjectives or short phrases given by users on Mechanical Turk.
The project is to provide code that programmatically groups the tags for each image url into 10-20 clusters, groups of tags that are semantically related to each other (synonyms, or closely related topics).
I've attached my attempt at this which is a python program that uses Wordnet, a crawl of Thesaurus.net, and a simple edit distance calculation to attempt to score the words relationships. This can be used as a starting point, or simply ignored if you have a better approach.
Sign in or Register to see more