Utvärdering av Amazon Machine Learning för taggsystem

Detta är en Kandidat-uppsats från KTH/Skolan för datavetenskap och kommunikation (CSC)

Författare: Farzana Madosh; Erik Lundsten; [2017]

Nyckelord: amazon; maskininlärning; taggsystem;

Sammanfattning: How companies deal with machine learning is currently a highly-discussed topic, as it can facilitate corporate manual work by training computers to recognize patterns and thus automate the working procedure. However, this requires resources and knowledge in the field. As a result, various companies like Amazon and Google provide machine learning services without requiring the user to have deep knowledge in the area. This study evaluates Amazon Machine Learning program for a tag system with data from the media company Newstag. In order to make this evaluation, a larger amount of data with tags was obtained from the company. The result of the study indicates that Amazon's program do not work for multilabel classification.

  HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)