DFKI SLT at GermEval 2021: Multilingual Pre-training and Data Augmentation for the Classification of Toxicity in Social Media Comments.
Rémi CalizzanoMalte OstendorffGeorg RehmPublished in: GermEval@KONVENS (2021)
Keyphrases
- data collection
- data sets
- social media
- training set
- image data
- database
- data analysis
- data points
- labelled data
- original data
- raw data
- classification accuracy
- pattern recognition
- big data
- high dimensional data
- data processing
- image classification
- supervised learning
- training data
- data sources
- data structure
- xml documents
- digital libraries
- feature extraction
- multi class
- data management
- decision trees
- training samples
- support vector machine svm
- training examples
- feature vectors
- support vector
- association rules
- social media data