Naive Bayes Spam Filtering Using Word-Position-Based Attributes.
Johan HovoldPublished in: CEAS (2005)
Keyphrases
- naive bayes
- spam filtering
- text classification
- attribute dependencies
- naive bayes classifier
- term frequency
- decision trees
- logistic regression
- text categorization
- feature selection
- classification algorithm
- classification accuracy
- probability estimation
- anti spam
- nominal attributes
- text classifiers
- uci datasets
- naive bayesian classifier
- machine learning
- uci data sets
- text mining
- attribute values
- labeled data
- bayesian network classifiers
- cost sensitive
- bayesian networks
- co occurrence
- test instances
- bayesian classifier
- spam filters
- training data
- conditional independence assumption
- continuous attributes
- naive bayes classification
- averaged one dependence estimators
- information gain
- term weighting
- unlabeled data
- data sets
- k nearest neighbor
- knn
- augmented naive bayes