Text Categorization Based on Boosting Association Rules

Yoon, Yongwook; Lee, G.G.

doi:10.1109/icsc.2008.70

Cited by 22 publications

(23 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Another improvement, Classification based on Predictive Association Rules (CPAR), assigns weights to the observations and upon correct classification, the weight of the observation is reduced [29]. CPAR can be viewed as a precursor to boosting PARs into a classifier model, which was proposed by Yoon et al [30]. Ozgur et al [22] proposed a procedure similar to growing decision trees on association rules for regression problems and Wang et al [27] grew decision trees on PARs with minimum support of 0.…”

Section: Background and Related Workmentioning

confidence: 98%

“…[21,20,29,30]), which is partly due to the ease of interpretation of the individual rules and partly due its good performance. Associative classification is less greedy than most other rule-based classifiers, because associative classification enumerates all PARs, while rulebased classifiers [7] or trees [4] consider only a subset of the PARs.…”

Section: Introductionmentioning

confidence: 97%

See 1 more Smart Citation

A simple statistical model and association rule filtering for classification

Simon

Kumar

2011

Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

Associative classification is a predictive modeling technique that constructs a classifier based on class association rules (also known as predictive association rules; PARs). PARs are association rules where the consequence of the rule is a class label. Associative classification has gained substantial research attention because it successfully joins the benefits of association rule mining with classification. These benefits include the inherent ability of association rule mining to extract high-order interactions among the predictors-an ability that many modern classifiers lack-and also the natural interpretability of the individual PARs.Associative classification is not without its caveats. Association rule mining often discovers a combinatorially large number of association rules, eroding the interpretability of the rule set. Extensive effort has been directed towards developing interestingness measures, which filter (predictive) association rules after they have been generated. These interestingness measures, albeit very successful at selecting interesting rules, lack two features that are highly valuable in the context of classification. First, only few of the interestingness measures are rooted in a statistical model. Given the distinction between a training and a test data set in the classification setting, the ability to make statistical inferences about the performance of the predictive classification rules on the test set is highly desirable. Second, the unfiltered set of predictive assocation rules (PARs) are often redundant, we can prove that certain PARs will not be used to construct a classification model given the presence of other PARs.In this paper, we propose a simple statistical model towards making inferences on the test set about the various performance metrics of predictive association rules. We also derive three filtering criteria based on hypothesis testing, which are very selective (reduce the number of PARs to be considered by the classifier by several orders of magnitude), yet do not effect the performance of the classification adversely. In the case, where the classification model is constructed as a logistic model on top of the PARs, we can mathematically prove, that the filtering criteria do not significantly effect the classifier's performance. We also demonstrate empirically on three publicly available data sets that the vast reduction in the number of PARs indeed did not come at the cost of reducing the predictive performance.

show abstract

Section: Background and Related Workmentioning

confidence: 98%

Section: Introductionmentioning

confidence: 97%

A simple statistical model and association rule filtering for classification

Simon

Kumar

2011

Proceedings of the 17th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining

View full text Add to dashboard Cite

show abstract

“…Table 4 depicts a comparison results between the proposed algorithm against other well-known classifiers. It should be noted that the results of the BCAR algorithm is reported in (Yoon and Lee, 2008) while for MCAR the results were obtained via experiment. Table 4 reveals that the proposed method has the highest accuracy for three of the dataset-Acq, Grain and Money-FX.…”

Section: Experiments and Resultsmentioning

confidence: 99%

“…There are many AC algorithms disseminated in literature in the last decade, such as CBA (Liu et al, 1999), CMAR (Li et al, 2001), CPAR (Han, 2003), MCAR (Thabtah et al, 2005), CACA (Tang and Liao, 2007), ACCF (Li et al, 2008), BCAR (Yoon and Lee, 2008), MAC (Abdelhamid et al, 2012) and CBAR (Han, 2003) PCBA (Chen et al, 2012). These techniques use different methods to discover the rules, sort the rules, store rules, filter out redundant rules and assigns the right class to test cases.…”

Section: Introductionmentioning

confidence: 99%

Partial Rule Match for Filtering Rules in Associative Classification

Refai¹

2014

Journal of Computer Science

View full text Add to dashboard Cite

In this study, we propose a new method to enhance the accuracy of Modified Multi-class Classification based on Association Rule (MMCAR) classifier. We introduce a Partial Rule Match Filtering (PRMF) method that allows a minimal match of the items in the rule's body in order for the rule to be added into a classifier. Experiments on Reuters-21578 data sets are performed in order to evaluate the effectiveness of PRMF in MMCAR. Results show that the MMCAR classifier performs better as compared to the chosen competitors.

show abstract

“…When applying association rules to document classification, the focus is on searching for the association of words that appear concurrently in a document. Yoon and Lee [26] used association rules for analyzing the document sets of Reuters-21578 and 20 newsgroups. They searched for rules from the training document, and then with an appropriate screener, they set up a rule database.…”

Section: Document Classificationmentioning

confidence: 99%