Locally Weighted Naive Bayes

By Eibe Frank, Mark A. Hall, Bernhard Pfahringer. Conference on Uncertainty in Artificial Intelligence, 2002

381 citations recorded by the source corpus.

Despite its simplicity, the naive Bayes classifier has surprised machine learning researchers by exhibiting good performance on a variety of learning problems. Encouraged by these results, researchers have looked to overcome naive Bayes' primary weakness-attribute independence-and improve the performance of the algorithm. This paper presents a locally weighted version of naive Bayes that relaxes the independence assumption by learning local models at prediction time. Experimental results show that locally weighted naive Bayes rarely degrades accuracy compared to standard naive Bayes and, in many cases, improves accuracy dramatically. The main advantage of this method compared to other techniques for enhancing naive Bayes is its conceptual and computational simplicity.

Subfields: Artificial Intelligence, Computational Theory and Mathematics

View Paper: https://www.semanticscholar.org/paper/d318c093cb450233dddef5e52833c9371860dc29

How researchers see this paper

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

In Artificial Intelligence

In Computational Theory and Mathematics

Citations in the corpus

Naive Bayes: applications, variations and vulnerabilities: a review of literature with code snippets for implementation (2020, methodology): “In another study, Frank et al. (2002) use a locally weighted version of NB by relaxing the independence assumption. According to their experimental results, the proposed method performs better than the NB in empirically for the data sets they considered. Furthermore, they show this method exhibits r”
Online Defect Prediction for Imbalanced Data (2015, methodology): “We experiment with the following updatable learning algorithms: Bayes [31], IBK [32], KStar [33], LogitBoost [34], LWL [35], NNge [36], and SPegasos [37].”
Multinomial Naive Bayes for Text Categorization Revisited (2004, methodology): “Note that, unlike [13] we did not normalize the feature values to lie in [0, 1] as we found it to degrade performance.”
Do we need hundreds of classifiers to solve real world classification problems? (2014, background): “LWL w, Local Weighted Learning (Frank et al., 2003), is an ensemble of DecisionStump base classifiers.”
A network intrusion detection system based on a Hidden Naïve Bayes multiclass classifier (2012, methodology): “NBTree (Kohavi, 1996) and DTNB (Hall, 2008) models combine the Naïve Bayes and decision tables, whereas the locally weighted Naïve Bayes (LWNB) (Frank, Hall, & Pfahringer, 2003), lazy Bayesian rule (LBR) and selective neighborhood Naïve Bayes (SNNB) (Xie, Hsu, Zongtian, & Lee, 2002) models are…”
A Novel Bayes Model: Hidden Naive Bayes (2009, methodology): “[23] presented an algorithm, called locally weighted naive Bayes, simply LWNB.”
Graph-based approach for airborne light detection and ranging segmentation (2014, background): “LWL w, Local Weighted Learning (Frank et al., 2003), is an ensemble of DecisionStump base classifiers.”
Identification of fraudulent financial statements using linguistic credibility analysis (2011, methodology): “Locally weighted learning (LWL) algorithm was used with the following parameters: use all neighbors in width of weighting function, classifier is decision stump, no normalization, and linear weighting kernel [2,22].”
Transfer learning for cross-company software defect prediction (2012, neutral): “Give a test instance x, by using sample weighting method described in [33], the conditional probability for jth attribute a j is Pða j jcÞ ¼ P n i¼1 w i dða ij ; a j Þdðc i ; cÞ þ 1 P n i¼1 w i dðc i ; cÞ þ n j ð8Þ where a ij is the value of jth attribute in ith training instance, and n j is the…”