Connecting the Dots: What Graph-Based Text Representations Work Best for Text Classification using Graph Neural Networks?
Margarita BugueñoGerard de MeloPublished in: EMNLP (Findings) (2023)
Keyphrases
- text classification
- neural network
- text data
- text documents
- text mining
- graph model
- text classifiers
- graph construction
- knn
- graph theoretic
- graph based algorithm
- label propagation
- text categorization
- k nearest neighbor
- graph theoretical
- bag of words
- labeled data
- training corpus
- graph representation
- naive bayes
- graph theory
- directed graph
- semantic representations
- fuzzy logic
- information retrieval
- multi label
- n gram
- random walk
- pattern recognition
- graph based semi supervised learning
- data cleaning
- graph structure
- protein function prediction
- artificial neural networks
- learning algorithm
- vector representation
- black and white
- graph connectivity
- connected components
- normalized cut
- graph partitioning
- semantic features
- weighted graph
- semi supervised learning
- knowledge discovery
- feature selection
- back propagation
- web documents
- unlabeled data
- semi supervised
- automatically discovering
- genetic algorithm
- automatic text classification