Adapted One-versus-All Decision Trees for Data Stream Classification

Hashemi, Sattar; Yang, Ying; Mirzamomen, Zahra; Kangavari, Mohammad Reza

doi:10.1109/tkde.2008.181

Cited by 80 publications

(32 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To compare the performance between OVO and OVA, studies such as [25] show that although there is no significant difference found between OVA and OVO, both strategies outperform the original classifier. In another study, the authors of [29] showed that OVA attains more accurate classification when comparing: (i) a concept-adapting very fast decision tree (CVFDT), a single multi-classifier; (ii) a weighted classifier ensemble (WCE), and streaming ensemble algorithm (SEA), both of which are ensembles of multi-class classifiers; and (iii) an ultrafast forest tree (UFFT), an OVO method. In [30] it was shown that OVO outperformed OVA, in contrast to [27] where it was suggested that OVA performed as well as OVO.…”

Section: Introductionmentioning

confidence: 99%

Maximizing the Diversity of Ensemble Random Forests for Tree Genera Classification Using High Density LiDAR Data

Sohn

Remmel

et al. 2016

Remote Sensing

View full text Add to dashboard Cite

Recent research into improving the effectiveness of forest inventory management using airborne LiDAR data has focused on developing advanced theories in data analytics. Furthermore, supervised learning as a predictive model for classifying tree genera (and species, where possible) has been gaining popularity in order to minimize this labor-intensive task. However, bottlenecks remain that hinder the immediate adoption of supervised learning methods. With supervised classification, training samples are required for learning the parameters that govern the performance of a classifier, yet the selection of training data is often subjective and the quality of such samples is critically important. For LiDAR scanning in forest environments, the quantification of data quality is somewhat abstract, normally referring to some metric related to the completeness of individual tree crowns; however, this is not an issue that has received much attention in the literature. Intuitively the choice of training samples having varying quality will affect classification accuracy. In this paper a Diversity Index (DI) is proposed that characterizes the diversity of data quality (Q i ) among selected training samples required for constructing a classification model of tree genera. The training sample is diversified in terms of data quality as opposed to the number of samples per class. The diversified training sample allows the classifier to better learn the positive and negative instances and; therefore; has a higher classification accuracy in discriminating the "unknown" class samples from the "known" samples. Our algorithm is implemented within the Random Forests base classifiers with six derived geometric features from LiDAR data. The training sample contains three tree genera (pine; poplar; and maple) and the validation samples contains four labels (pine; poplar; maple; and "unknown"). Classification accuracy improved from 72.8%; when training samples were selected randomly (with stratified sample size); to 93.8%; when samples were selected with additional criteria; and from 88.4% to 93.8% when an ensemble method was used.

show abstract

Section: Introductionmentioning

confidence: 99%

Maximizing the Diversity of Ensemble Random Forests for Tree Genera Classification Using High Density LiDAR Data

Sohn

Remmel

et al. 2016

Remote Sensing

View full text Add to dashboard Cite

show abstract

“…They are based on an underlying assumption that the training data set does not contain any uncertainty information and properly represents the examined concept. However, in many real-world applications, data change its nature over time-which is a vital problem for data stream analysis (Hashemi et al 2009). For example, in environmental monitoring applications, data may change according to the examined conditions and what once was considered an outlier may in near future become a representative of the target concept.…”

Section: One-class Classificationmentioning

confidence: 99%

One-class classifiers with incremental learning and forgetting for data streams with concept drift

Krawczyk

Woźniak

2014

Soft Comput

View full text Add to dashboard Cite

One of the most important challenges for machine learning community is to develop efficient classifiers which are able to cope with data streams, especially with the presence of the so-called concept drift. This phenomenon is responsible for the change of classification task characteristics, and poses a challenge for the learning model to adapt itself to the current state of the environment. So there is a strong belief that one-class classification is a promising research direction for data stream analysis-it can be used for binary classification without an access to counterexamples, decomposing a multi-class data stream, outlier detection or novel class recognition. This paper reports a novel modification of weighted one-class support vector machine, adapted to the non-stationary streaming data analysis. Our proposition can deal with the gradual concept drift, as the introduced one-class classifier model can adapt its decision boundary to new, incoming data and additionally employs a forgetting mechanism which boosts the ability of the classifier to follow the model changes. In this work, we propose several different strategies for incremental learning and forgetting, and additionally we evaluate them on the basis of several real data streams. Obtained results confirmed the usability of proposed classifier to the problem of data stream classification with the presence of concept drift. Additionally, implemented

show abstract

“…The novel class recognition part is dogged by building it more adaptive to the developing stream and allowing it to notice more than one new class at a time. The paper [11] supported data stream categorization. At first, there is low error association and therefore high variety amongst component classifiers, which tends to elevated categorization correctness.…”

Section: Literature Reviewmentioning

confidence: 99%

Bucketization based Flow Classification Algorithm for Data Stream Privacy Mining

Kesavaraj¹,

Sukumaran²

2013

IJCA

View full text Add to dashboard Cite

In recent years, data mining plays a major role in maintaining the huge volume of data from which it can derive the useful information. With the huge number of formation of data, the data wants to be lectured in a limit to the charge of growth. But it is complex to get over the set of meaningful information from the continuous set of data. Data-stream mining is a method which can discover important information from a huge contract of prehistoric data. For identification of useful information, the classification of continuous data streams is done. Current approaches in classifying the data streams are processed using supervised learning algorithms, which can be qualified with tagged data. Usually, manual classification of data is both expensive and time consuming. As a result, where massive amount of data emerge at a high speed, tagged data might be very sparse. Therefore, only a restricted amount of training data might be accessible for constructing the classification models, tend to badly trained classifiers. To overcome the issue, in this work, a novel technique is presented to build a classification set having both unlabeled and a small amount of labeled instances. This model is built by using the Flow Classification Algorithm (FCA). The FC algorithm is able to judge internally on set of marked data. Before classification, the correlation set of attributes in the each record set are grouped using bucketization technique. The superiority of models updated from them is enough for utilization of unlabeled records, or whether more set of labeled records are needed for classification is processed. Experimental evalaution is conducted to the proposed FC technqiue over its counterparts to find a set of diverse solution in terms of execution time, classification accuracy and security. Performance metrics for evaluation of proposed FCA technique shows that the security level is 10-15% high against existing work.

show abstract

Adapted One-versus-All Decision Trees for Data Stream Classification

Cited by 80 publications

References 19 publications

Maximizing the Diversity of Ensemble Random Forests for Tree Genera Classification Using High Density LiDAR Data

Maximizing the Diversity of Ensemble Random Forests for Tree Genera Classification Using High Density LiDAR Data

One-class classifiers with incremental learning and forgetting for data streams with concept drift

Bucketization based Flow Classification Algorithm for Data Stream Privacy Mining

Contact Info

Product

Resources

About