Word2Vec: Optimal hyperparameters and their impact on natural language processing downstream tasks.
Tosin P. AdewumiFoteini LiwickiMarcus LiwickiPublished in: Open Comput. Sci. (2022)
Keyphrases
- hyperparameters
- natural language processing
- closed form
- model selection
- tasks in natural language processing
- word sense disambiguation
- prior information
- cross validation
- natural language
- bayesian framework
- optimal solution
- maximum likelihood
- machine learning
- bayesian inference
- gaussian process
- worst case
- incremental learning
- em algorithm
- information extraction
- incomplete data
- regularization parameter
- sample size
- maximum a posteriori
- noise level
- support vector
- bayesian networks
- missing values
- co occurrence
- support vector machine
- dynamic programming
- random sampling
- training data
- clustering algorithm
- learning algorithm