Humor Detection in Product Question Answering Systems Dataset Questions extracted from the Amazon website. Please cite: Humor Detection in Product Question Answering Systems. Yftah Ziser, Elad Kravi & David Carmel. SIGIR 2020. Authors: - Yftah Ziser (yftahz@amazon.com) - Elad Kravi (ekravi@amazon.com) - David Carmel (dacarmel@amazon.com) Dataset Structure: The dataset contains 3 csv file where each row corresponds to a product question. Humorous.csv - contains the humorous product questions, link for direct download - https://humor-detection-pds.s3-us-west-2.amazonaws.com/Humorous.csv Non-humorous-unbiased.csv - contains the non-humorous prodcut questions, from the same products as the humorous one, link for direct download - https://humor-detection-pds.s3-us-west-2.amazonaws.com/Non-humorous-unbiased.csv Non-humorous-biased.csv - contains the non-humorous prodcut questions, from randomly selected products, link for direct download - https://humor-detection-pds.s3-us-west-2.amazonaws.com/Non-humours-biased.csv All files have the same structure The columns are: - question: the question text - product_description: short description of the product. - image_url: url for the prodcut image. - label: 1 if the product question is humorous, 0 otherwise.