IRAHC: Instance Reduction Algorithm using Hyperrectangle Clustering. Issue 5 (May 2015)

• Record Type: Journal Article
• Authors: Hamidzadeh, Javad; Monsefi, Reza; Sadoghi Yazdi, Hadi
• Is Part Of: Pattern recognition. Volume 48:Issue 5(2015:May)
• Page Start: 1878
• Page End: 1889
• Publication Date: 2015-05
• Subjects: Instance reduction -- Instance selection -- Hyperrectangle -- Instance-based classifiers -- k-Nearest neighbor (k-NN); Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2014.11.005
• Languages: English
• ISSNs: 0031-3203

Abstract

Abstract: In instance-based classifiers, there is a need for storing a large number of samples as training set. In this work, we propose an instance reduction method based on hyperrectangle clustering, called Instance Reduction Algorithm using Hyperrectangle Clustering (IRAHC). IRAHC removes non-border (interior) instances and keeps border and near border ones. This paper presents an instance reduction process based on hyperrectangle clustering. A hyperrectangle is an n-dimensional rectangle with axes aligned sides, which is defined by min and max points and a corresponding distance function. The min–max points are determined by using the hyperrectangle clustering algorithm. Instance-based learning algorithms are often confronted with the problem of deciding which instances must be stored to be used during an actual test. Storing too many instances can result in a large memory requirements and a slow execution speed. In IRAHC, core of instance reduction process is based on set of hyperrectangles. The performance has been evaluated on real world data sets from UCI repository by the 10-fold cross-validation method. The results of the experiments have been compared with state-of-the-art methods, which show superiority of the proposed method in terms of classification accuracy and reduction percentage. Highlights: An instance reduction method has been proposed by using hyperrectangle clustering. The performance of the proposed method has been examined on real data sets. The results have been compared to seven important instance reduction algorithms. The proposed method yields the lowest classification error rate significantly. The proposed method has the best instance reduction percentage significantly.