Xavier Holt edited Algorithm_Solution_1_page_Baseline__.md  almost 8 years ago

Commit id: d3edaf09bb814e40eb8a06bb5d66b5363c6f6b05

deletions | additions      

       

* Output: per-article tag of whether the article should be included in a brief/confidence of brief-worthiness.  * Application: Search for an entity, find the k- most ‘brief-worthy’ articles, perform sentence selection.  * Training data: Hugo app, ~5000 brief queries with linked articles and article-level indicator of whether they were included in final brief.  ### Features  * nEn: distinct named entities in article.  * C: Vector of named entity mentions weighted by confidence.  * FK: Sentence complexity.  * C: Vector of named entity mentions weighted by confidence.