Hybrid decision tree and naïve Bayes classifiers for multi-class classification tasks

Farid, Dewan Md.; Zhang, Li; Rahman, Chowdhury Mofizur; Hossain, Mohammad Alamgir; Strachan, Rebecca

doi:10.1016/j.eswa.2013.08.089

Cited by 348 publications

(134 citation statements)

References 28 publications

Supporting

Mentioning

133

Contrasting

Unclassified

Order By: Relevance

“…· C4.5, standard decision tree. · NBTree [19], decision tree with naive Bayes as the leaf node. · RTAN [20], tree-augmented naive Bayes ensembles.…”

Section: Statistical Results On Uci Data Setsmentioning

confidence: 99%

How to Mine Information from Each Instance to Extract an Abbreviated and Credible Logical Rule

Wang

Sun

Cao

2014

Entropy

View full text Add to dashboard Cite

Decision trees are particularly promising in symbolic representation and reasoning due to their comprehensible nature, which resembles the hierarchical process of human decision making. However, their drawbacks, caused by the single-tree structure, cannot be ignored. A rigid decision path may cause the majority class to overwhelm other class when dealing with imbalanced data sets, and pruning removes not only superfluous nodes, but also subtrees. The proposed learning algorithm, flexible hybrid decision forest (FHDF), mines information implicated in each instance to form logical rules on the basis of a chain rule of local mutual information, then forms different decision tree structures and decision forests later. The most credible decision path from the decision forest can be selected to make a prediction. Furthermore, functional dependencies (FDs), which are extracted from the whole data set based on association rule analysis, perform embedded attribute selection to remove nodes rather than subtrees, thus helping to achieve different levels of knowledge representation and improve model comprehension in the framework of semi-supervised learning. Naive Bayes replaces the leaf nodes at the bottom of the tree hierarchy, where the conditional independence assumption may hold. This technique reduces the potential for overfitting and overtraining and improves the prediction quality and generalization. Experimental results on UCI data sets demonstrate the efficacy of the proposed approach.

show abstract

“…· C4.5, standard decision tree. · NBTree [19], decision tree with naive Bayes as the leaf node. · RTAN [20], tree-augmented naive Bayes ensembles.…”

Section: Statistical Results On Uci Data Setsmentioning

confidence: 99%

How to Mine Information from Each Instance to Extract an Abbreviated and Credible Logical Rule

Wang

Sun

Cao

2014

Entropy

View full text Add to dashboard Cite

show abstract

“…Classification is a supervised learning method intended to build a detection model which could explain a data class, sort out data points or elements into classes that correspond to their features based on a sequence of pre-determined class label [13], [22]. Consequently, the structured label elements will be further used to predict the class label of new elements.…”

Section: IImentioning

confidence: 99%

“…Thus, clustering groups objects based on characterization of data points, where every single data point in a cluster is alike to those within its cluster, but different from those in different or other clusters [20], [21]. Clustering has capabilities to group similar malicious data points collectively into one or more cluster including previously unseen malicious data points [22]. In addition, k-means clustering is much more efficient as a contrast to other existing clustering methods as the fact that this algorithm able to process huge volumes of instances quickly and have nonlinear complexity.…”

Section: IImentioning

confidence: 99%

An Improved Malicious Behaviour Detection Via k-Means and Decision Tree

Yassin¹,

Rahayu²,

Abdollah³

et al. 2016

ijacsa

View full text Add to dashboard Cite

Abstract-DataMining algorithm which is applied as an anomaly detection system has been considered as one of the essential techniques in malicious behaviour detection. Unfortunately, such detection system is known for its inclination in detecting a cyber-malicious activity more accurately (i.e. maximizing malicious and non-malicious behaviours detection) and has become a persistent limitation in the deployment of intrusion detection systems. Consequently, these constraints will affect a number of important performance factors such as the accuracy, detection rate and false alarms. In this research, KMDT proposed as an anomaly detection model that utilized kmeans clustering and decision tree classifier to maximize the detection of malicious behaviours by scrutinizing packet headers. The k-means clustering employed for labelling and plots the whole behaviours into identical cluster, which characterized the behaviours into suspicious or non-suspicious composition. Subsequently, these dissimilar clustered behaviours are reordered within two classes of types such as malicious and nonmalicious via decision tree classifier. KMDT is a profitable finding which improved the anomaly detection performance in identifying suspicious and non-suspicious behaviours as well as characterizes it into malicious and non-malicious behaviours more accurately. These criteria have been validated by the result from the experiments throughout banking system environment dataset 2016. KMDT have detected more malicious behaviours accurately as contrast to discrete and diversely combined methods.

show abstract

“…Decision trees are effectively applied when the number of object features is not great [7], [8]. Except random forests [9], boosting is applicable when the number of object features is in the order of hundreds or thousands [10], [11].…”

Section: Related Workmentioning

confidence: 99%

Optimal Training Parameters and Hidden Layer Neuron Number of Two-Layer Perceptron for Generalised Scaled Object Classification Problem

Romanuke

2015

Information Technology and Management Science

View full text Add to dashboard Cite

-The research is focused on optimising two-layer perceptron for generalised scaled object classification problem. The optimisation criterion is minimisation of inaccuracy. The inaccuracy depends on training parameters and hidden layer neuron number. After its statistics is accumulated, minimisation is executed by a numerical search. Perceptron is optimised additionally by extra training. As it is done, the classification error percentage does not exceed 3 % in case of the worst scale distortion.

show abstract

Hybrid decision tree and naïve Bayes classifiers for multi-class classification tasks

Cited by 348 publications

References 28 publications

How to Mine Information from Each Instance to Extract an Abbreviated and Credible Logical Rule

How to Mine Information from Each Instance to Extract an Abbreviated and Credible Logical Rule

An Improved Malicious Behaviour Detection Via k-Means and Decision Tree

Optimal Training Parameters and Hidden Layer Neuron Number of Two-Layer Perceptron for Generalised Scaled Object Classification Problem

Contact Info

Product

Resources

About