Zheng, Z., and Low, B.T. Classifying unseen cases with many missing values. In Proceedings of the 3 ra Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD-99), Springer-Verlag, 1999.  Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
MRDTL: A multi-relational decision tree learning algorithm - Leiva (2002) (1 citation) (Correct)
....are likely to be present (Liu et al. 1997) and most of the time they are not representative of any subgroup of instances. There has been substantial research done on handling missing values in machine learning (Grzymala Busse and Hu, 2001, Liu et al. 1997, Ortega and Numao, 1999, Quinlan, 1989, Zheng and Low, 1999) Several of them provide a comparison between different methods of dealing with missing or unknown values (e.g. Quinlan, 1989) but others like (Ortega and Numao, 1999) propose new or improved methods. One of the first approaches to tackle this problem was to ignore those instances with ....
Zheng, Z., and Low, B.T. Classifying unseen cases with many missing values. In Proceedings of the 3 ra Pacific-Asia Conference on Knowledge Discovery and Data Mining (PAKDD-99), Springer-Verlag, 1999.
Classifying Unseen Cases with Many Missing Values - Zheng, Low (1999) (1 citation) Self-citation (Zheng Low) (Correct)
....Section 3 empirically explores the robustness of C4.5 and the four committee learning algorithms for tolerating missing attribute values in test data. The last section summaries conclusions and outlines some directions for future research. The full version of this paper is available from [10], which also contains an approach to improving the robustness of the committee learning techniques for tolerating missing attribute values in test data. 2 Boosting, Bagging, SASC, and SASCMB Bagging [3] generates different classifiers using different bootstrap samples. Boosting [2, 4, 7] builds ....
....Figure 1 shows the average error rates of the five learning algorithms over the 37 domains as a function of the missing attribute value level L. The detailed error rates of these algorithms and error ratios of each committee learning algorithm over C4.5 and more discussions can be found in [10]. From these experimental results, we have the following observations. 1) All the four committee learning algorithms can significantly reduce the average error of C4.5 at all missing attribute value level from 0 to 50 across the 37 domains. Among them, SascMB always has the lowest average ....
Zheng, Z. and Low, B.T.: Classifying unseen cases with many missing values. Tech Report (TR C99/02) (available at http://www3.cm.deakin.edu.au/~zijian/ Papers/comm-missing-trC99-02.ps.gz), School of Computing and Mathematics, Deakin University, Australia (1999).
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC