| Autor: |
Hammarström, Harald, Rönchen, Philipp, Elgh, Erik, Wiklund, Tilo |
| Předmět: |
|
| Zdroj: |
Theoretical Linguistics; Oct2019, Vol. 45 Issue 3/4, p233-245, 13p, 1 Diagram, 1 Graph |
| Abstrakt: |
We welcome Gerhard Jäger's framing of Computational Historical Linguistics: its history and background, its goals and ambitions as well as the concrete implementation by Jäger himself. In his case study, Jäger himself is using Machine Learning algorithms (Sections 3.2 and 3.6), but also performing statistical hypothesis testing (Section 3.1) and phylogenetic inference (Sections 3.4 and 3.5). The section on cognate clustering is concluded with the claim that "Based on evaluations against manually assembled cognacy judgments for different but similar data (Jäger and Sofroniev, 2016; Jäger et al., 2017), we can expect an average I F i -score of 60-80% for automatic cognate detection". Here, in the position paper by Jäger, one would have expected accuracy numbers on cognate detection using the method actually presented for the data on Romance (or at least a subfamily of similar depth). [Extracted from the article] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
