| ID |
Question |
Name |
| Jiao Concept Mapping GTW1 |
The authors claim that "similarity features can be chosen to be language-independent." Do they test this claim? The example fields given (ISBN and dates) seem like they would provide very little information (dates), or would only be applicable in specific cases where the medium in question has unique identifiers (ISBN) and is used with both thesauri. Are there enough useful language-independent features in general to make this a useful approach? |
Gregory Todd Williams |
| Jiao Concept Mapping Jesse Weaver |
The authors often mention or allude to how this approach could be used in a multi-lingual setting (sections 1, 3.1, 4.4, and 5). As stated in the conclusion, further investigation of this claim is left to future work. How do you foresee this approach being useful in a multi-lingual setting? |
Jesse Weaver |
| Jiao Concept Mapping Joshua Shinavier 1 |
To what extent did the choice of a classifier technique (Markov networks) affect the results of the various experiments? Would it have been possible to apply another variety of classifier (e.g. support vector machines) with better results for some experiments? |
Joshua Shinavier |