Spam Filtering with Naive Bayes - Which Naive Bayes?

By Vangelis Metsis, Ion Androutsopoulos, G. Paliouras. 2006

707 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Certified Defenses for Data Poisoning Attacks (2017, methodology): “We ran both the upper bound relaxation and the IQP solver on two text datasets, the Enron spam corpus (Metsis et al., 2006) and the IMDB sentiment corpus (Maas et al., 2011).”
Data Shapley: Equitable Valuation of Data for Machine Learning (2019, methodology): “In the ﬁrst exper-iment, we use the spam classiﬁcation data set (Metsis et al., 2006).”
Machine learning for Internet of Things data analysis: A survey (2017, methodology): “Moreover, they are a popular model for applications such as spam filtering [97], text categorization, and automatic medical diagnosis [98].”
Adversarial Attacks on Deep Learning Models in Natural Language Processing: A Survey (2019, methodology): “[34] performed spam detection on Enron Spam Dataset [90] and adopted six large datasets from [157], i.e., AG’s news12, Sogou news [139], DBPedia ontology dataset, Yahoo!”
A review of machine learning approaches to Spam filtering (2009, methodology): “Metsis et al. (2006) compared the performance of five naive Bayes classifiers.”
Red Alarm for Pre-trained Models: Universal Vulnerabilities by Neuron-Level Backdoor Attacks (2021, methodology): “For the evaluation of NLP PTMs, we use SST-2 (Socher et al., 2013), OLID (Zampieri et al., 2019), and Enron (Metsis et al., 2006).”
A survey of the applications of text mining in financial domain (2016, methodology): “Various Studies on the Spam Detection Study Dataset Model Performance Measure Zhan et al. (2011) SpamAssasin NB Precision, Recall, ROC Chen et al. (2008) PU Corpus AODE/ HNB/ LWNB/ NB/ SVM/ C 4.5/ k-NN Accuracy Brodsky and Brodsky (2007) SpamAssassin Trinity Precision Metsis et al. (2006) SpamAssass”
Black-Box Generation of Adversarial Text Sequences to Evade Deep Learning Classifiers (2018, methodology): “• Datasets: In our experiments, we use 7 large-scale text datasets from [30] together with the Enron Spam Dataset [15], which includes a variety of NLP tasks, e.”
Interpretable Policies for Dynamic Product Recommendations (2016, methodology): “This problem arises in many domains such as recommender systems [21, 6] and personalized online advertising [8, 16, 7]; similar methods for matching customers to products are even used for email classification and spam detection [20].”
Using probabilistic generative models for ranking risks of Android apps (2012, methodology): “Machine Learning in Security: Naive Bayes has been extensively used both in the context of spam detection [25, 18, 16, 28] and anomaly detection [26, 4] in network traffic flows.”
Understanding Black-box Predictions via Influence Functions (2017, methodology): “These results are on the Enron1 spam dataset (Metsis et al., 2006), with 4,147 training and 1,035 test examples; we trained logistic regression on a bag-of-words representation of the emails.”
Scalable Optimization of Randomized Operational Decisions in Adversarial Classification Settings (2015, methodology): “of AAS as a spam filtering task to compare the classification accuracy with the state of the art alternatives [21, 8].”
Spam Filtering Using Inexact String Matching in Explicit Feature Space with On-Line Linear Classifiers (2006, methodology): “Traditional statistical spam filters are built on a foundation of Naive Bayes classification on word-level features (Graham, 2002; Metsis et al., 2006), a combination that has proven efficient and reasonably effective.”