Template-based pun extractor and generator
Computational Humor is a growing research field. Computers are being used to communicate with humans. And to make computer conversation more human and natural, humor is needed. Studying humor will give an insight of human creativity, different from simply having a controlled set of sentences (Binste...
Saved in:
| Main Author: | |
|---|---|
| Format: | text |
| Language: | English |
| Published: |
Animo Repository
2008
|
| Subjects: | |
| Online Access: | https://animorepository.dlsu.edu.ph/etd_masteral/3473 https://animorepository.dlsu.edu.ph/context/etd_masteral/article/10311/viewcontent/CDTG004251_P.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | De La Salle University |
| Language: | English |
| Summary: | Computational Humor is a growing research field. Computers are being used to communicate with humans. And to make computer conversation more human and natural, humor is needed. Studying humor will give an insight of human creativity, different from simply having a controlled set of sentences (Binsted, 2006).
Pun generator systems depend on manually created templates which require a lot of effort to build. This constrains the possible types of jokes that can be generated. The Template-based Pun Extractor and Generator is a system that learns templates based on training examples. It uses this learned knowledge to generate punning riddles. The learning algorithm utilizes semantic and phonetic knowledge to capture the wordplay used for a joke. Templates contain word relationships, variables and tags. The generation algorithm produces punning riddles based on user keywords and the learned templates. The use of keywords ensures that the joke will be relevant to what the user knows or is interested in. The system is evaluated based on analysis of the corpus, linguist evaluation, and user feedback. Linguist evaluation garnered an average score of 4.0 out of 5, as this evaluated the completeness of the learned templates. The user feedback for generated puns had a score of 2.13 out of 5, while the user feedback for human-made jokes is 2.70 out of 5. This user feedback is to check whether system-generated jokes are as funny as human-made jokes. |
|---|