metadata
license: mit
Published: September 21th, 2022
Author: Julius Breiholz
GARFAB-Dataset
The (G)erman corpus of annotated (A)pp (R)eviews to detect (F)eature requests (A)nd (B)ug reports (GARFAB) is a dataset to fine-tune models regarding classification of app reviews (ASRs) into "Feature Requests", "Bug Reports" and "Irrelevants" for the German language. All ASRs were collected from the Google Play Store and were classified manually by two independent annotators. A weighted and a full version are published with the following distributions of ASRs:
Feature Request | Bug Reports | Irrelevant | Total | |
---|---|---|---|---|
full | 345 | 387 | 2212 | 2944 |
weighted | 345 | 345 | 345 | 1035 |