nikrao opened a new pull request #7390: adding ranking metrics (precision/recall) at position K. URL: https://github.com/apache/incubator-mxnet/pull/7390 The file takes in data in dictionary format, for both predictions and ground truth. Ranking problems are very common in several areas of ML, and specifically computing metrics at a certain position is common in many applications. Usually for recommender systems, we don't want to compute the overall precision, but only, say precision@10, because only 10 items might be shown to the user. It makes sense to optimize your model that gets the top entries right, sacrificing overall precision. Similar arguments hold for recall, coverage etc. Having such a standard metric available in MXNet will let more people in the ML community make use of it, and increase it's adoptation. ---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org
With regards, Apache Git Services