The Optimality of Naive Bayes

By Harry Zhang. The Florida AI Research Society, 2004

1,994 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/0bf71f7b1fa5f95b50d27e3583c81ffe7178e58c

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

Neural Associative Memory with Optimal Bayesian Learning (2011, background): “Although this assumption is almost never fulfilled in real-world data, experiments reveal that naive Bayesian classifiers perform surprisingly well or even optimal in many domains that contain clear attribute dependencies (Zhang, 2004; Domingos & Pazzani, 1997).”
Optimal synaptic learning in non-linear associative memory (2010, methodology): “This paper develops the optimal neural associative mem- ory based on Bayesian maximum-likelihood considerations assuming local synaptic learning (section II) and indepen- dent address attributes [56], [57].”
Prediction of Slope Stability using Naive Bayes Classifier (2018, background): “Despite the naive design and the apparently oversimplified assumptions, many examples are reported exhibiting the surprising effectiveness of the NBCs even with strong dependencies (Hand and Yu, 2001; Zhang, 2004).”
A Survey on Human Activity Recognition using Wearable Sensors (2013, background): “The following values can be obtained from the confusion matrix in a binary classification problem: • True Positives (TP): The number of positive instances that were classified as positive.”
Machine learning for Internet of Things data analysis: A survey (2017, unknown): “Different naive Bayes classi ﬁ ers use different approaches and distributions to estimate p ð t ¼ c Þ and p ð z j t ¼ c Þ [61].”
Stochastic Simulation of Patterns Using Distance-Based Pattern Modeling (2010, background): “The process can be perceived as a naive Bayesian classifier, where different sources of information are fused together, and a class label is assigned to them (Anderson, 1974; Zhang, 2004).”
Disambiguation and co-authorship networks of the U.S. patent inventor database (1975–2010) (2014, methodology): “Despite its simplicity, Naïve Bayes classifiers erform surprisingly well in real problems (Hastie et al., 2001; ishop, 2006; Lewis, 1998; Rish, 2001; Zhang, 2004); the Author-ity pproach differs in that it does not require independence between ll attributes, only between the sets of attributes.”
Using data mining for bank direct marketing: an application of the CRISP-DM methodology (2011, methodology): “Examples of popular DM techniques are: Naïve Bayes (NB) (Zhang, 2004), Decision Trees (DT) (Aptéa and Weiss, 1997) and Support Vector Machines (SVM) (Cortes and Vapnik, 1995).”
Comparing Machine Learning Classifiers for Object-Based Land Cover Classification Using Very High Resolution Imagery (2014, methodology): “The NB classifier assumes that feature vectors from each land cover type are normally distributed, but not necessarily independently distributed, different from the other commonly-used classification model, naive Bayes [23,29].”
A Workflow for Visual Diagnostics of Binary Classifiers using Instance-Level Explanations (2017, methodology): “Training Test AUC Gaussian Na¨ıve Bayes (GNB) [33] 0.58 0.52 Logistic Regression (LR) [32] 0.85 0.79 Random Forest (RF) [6] 0.88 0.79 Multi Layer Perceptron (MLP) [14] 0.85 0.80 As we can see most of the models achieve similar performance on the test data.”
Bayesian Naïve Bayes classifiers to text classification (2018, background): “This paradox is explained by Zhang that true reason for its competitive performance in classification lies in the dependence distribution [10].”
Fraudulent Democracy? An Analysis of Argentina's Infamous Decade Using Supervised Machine Learning (2011, background): “This assumption largely simplifies the learning process, as the classifier is fully defined simply by the conditional probabilities of each attribute given the class (Zhang 2004).”
Using naive bayes to detect spammy names in social networks (2013, background): “The Naive Bayes algorithm has demonstrated great success in classiﬁcation problems (see e.g., [11, 24, 21]) even though the value s ( (cid:126)x ) of (2.3) is often highly inaccurate as an estimate of probability [4, 31, 10].”
dtControl: decision tree learning algorithms for controller representation (2020, methodology): “e considered various linear classification techniques including Logistic Regression [32, Chapter 4], linear Support Vector Machines (SVM) [32, Chapter 7], Perceptrons [32, Chapter 5], and Naive Bayes [33]. However, the latter two yielded significantly larger DTs in all of our experiments, so dtContr”
Comparative classification analysis of post-harvest growth detection from terrestrial LiDAR point clouds in precision agriculture (2015, background): “The advantage lies in the assumption of independent features, whereby only the variance of the features for each class label needs to be determined and not the entire covariance matrix (Zhang, 2004).”
Naive Bayes: applications, variations and vulnerabilities: a review of literature with code snippets for implementation (2020, background): “According to the studies (Hand and Yu 2001; Domingos and Pazzani 1996; Zhang 2004; Rish 2001), an interesting yet an important outcome suggests that the conditional independence is only a sufﬁcient condition for the optimal of the NB, but not a necessary condition.”
Efficient Water Quality Prediction Using Supervised Machine Learning (2019, methodology): “We used default values for the algorithms, except MLP, which uses a (3, 7) conﬁguration.”