Naive Bayesian Classifiers for Ranking

Abstract

Abstract. It is well-known that naive Bayes performs surprisingly well in classification, but its probability estimation is poor. In many applications, however, a ranking based on class probabilities is desired. For example, a ranking of customers in terms of the likelihood that they buy one's products is useful in direct marketing. What is the general performance of naive Bayes in ranking? In this paper, we study it by both empirical experiments and theoretical analysis. Our experiments show that naive Bayes outperforms C4.4, the most state-of-the-art decisiontree algorithm for ranking. We study two example problems that have been used in analyzing the performance of naive Bayes in classification [3]. Surprisingly, naive Bayes performs perfectly on them in ranking, even though it does not in classification. Finally, we present and prove a sufficient condition for the optimality of naive Bayes in ranking.

Classification accuracy performance of Naïve Bayesian (NB), Bayesian Networks (BN), Lazy Learning of Bayesian Rules (LBR) and Instance-Based Learner (IB1) - comparative study

“…However, none of these datasets were user profiles. If we consider our simulation results and previously mentioned outcomes of [18], [19] and [21], we can conclude that the Naïve Bayesian classifier performs well with different datasets including user profile dataset.…”

“…In [18], authors compared the ranking performance of NB classifier with DT (C4.4) classifier. The experiments conducted with using 15 dataset from UCI data repository [16] (see Table V).…”

Locally Weighted Learning: How and When Does it Work in Bayesian Networks?

“…ACC criterion has been successful used on many specific problems 24,22,23,41,39 . Nevertheless, in some data mining real world application, learning a classifier with accurate ranking or probability estimation is also desirable, not just only classification accuracy 45,38 . For example, in direct marketing, we often need to promote the top x% of customers during gradual roll-out, or we often deploy different promotion strategies to customers with different likelihood of buying some products.…”

Content Trust Model for Detecting Web Spam

“…Ranking SVM [14] is a method which formalizes learning to rank as learning for classification on pairs of instances and tackles the classification issue by using SVM. The reason why we use ranking SVM is because it performs best compare to the other method, such as Naive Bayesian [24] and decision tree [23] for ranking problem. The experiments result is described in section 4 lately.…”

“…We experimented with a variety of ranking techniques, and here we only present the following algorithms: decision-tree based ranking techniques (R-DT) [23], Naive Bayesian based ranker (R-NB) [24] and ranking support vector machine (R-SVM), which modified by us in section 3.2 to suit the problem of spam detection. All algorithms are implemented within the Weka framework [25].…”