Abstract
In this article, we propose a novel probabilistic framework to improve the accuracy of a weighted majority voting algorithm. In order to assign higher weights to the classifiers which can correctly classify hard-to-classify instances, we introduce the item response theory (IRT) framework to evaluate the samples’ difficulty and classifiers’ ability simultaneously. We assigned the weights to classifiers based on their abilities. Three models are created with different assumptions suitable for different cases. When making an inference, we keep a balance between the accuracy and complexity. In our experiment, all the base models are constructed by single trees via bootstrap. To explain the models, we illustrate how the IRT ensemble model constructs the classifying boundary. We also compare their performance with other widely used methods and show that our model performs well on 19 datasets.
Similar content being viewed by others
References
Z. H. Zhou. Ensemble learning. Encyclopedia of Biometrics, S. Z. Li, Ed., Berlin, Germany: Springer, pp. 411–416, 2009.
L. Lam, S. Y. Suen. Application of majority voting to pattern recognition: an analysis of its behavior and performance. IEEE Transactions on Systems, Man, and Cybernetics — Part A: Systems and Humans, vol. 27, no. 5, pp. 553–568, 1997. DOI: https://doi.org/10.1109/3468.618255.
A. F. R. Rahman, H. Alam, M. C. Fairhurst. Multiple classifier combination for character recognition: revisiting the majority voting system and its variations. In Proceedings of the 5th International Workshop on Document Analysis Systems, pp. 167–178, Springer, Princeton, USA, 2002.
H. Kim, H. Kim, H. Moon, H. Ahn. A weight-adjusted voting algorithm for ensembles of classifiers. Journal of the Korean Statistical Society, vol. 40, no. 4, pp. 437–449, 2011. DOI: https://doi.org/10.1016/j.jkss.2011.03.002.
S. E. Embretson, S. P. Reise. Item Response Theory, New York, USA: Psychology Press, 2013.
F. Martínez-Plumed, R. B. C. Prudencio, A. Martínez-Usó, J. Hernández-Orallo. Item response theory in AI: Analysing machine learning classifiers at the instance level. Artificial Intelligence, vol. 271, pp. 18–42, 2019. DOI: https://doi.org/10.1016/j.artint.2018.09.004.
L. Breiman. Bagging predictors. Machine Learning, vol. 24, no. 2, pp. 123–140, 1996. DOI: https://doi.org/10.1007/BF00058655.
I. Gandhi, M. Pandey. Hybrid ensemble of classifiers using voting. In Proceedings of International Conference on Green Computing and Internet of Things, IEEE, Noida, India, pp. 399–404, 2015. DOI: https://doi.org/10.1109/ICGCIoT.2015.7380496.
A. Rojarath, W. Songpan, C. Pong-Inwong. Improved ensemble learning for classification techniques based on majority voting. In Proceedings of the 7th IEEE International Conference on Software Engineering and Service Science, IEEE, Bei**g, China, pp. 107–110, 2016. DOI: https://doi.org/10.1109/ICSESS.2016.7883026.
C. Cornelio, M. Donini, A. Loreggia, M. S. Pini, F. Rossi. Voting with random classifiers (vorace). ar**v: 1909.08996, 2019. https://arxiv.org/abs/1909.08996.
X. B. Liu, Z. T. Liu, G. J. Wang, Z. H. Cai, H. Zhang. Ensemble transfer learning algorithm. IEEE Access, vol. 6, pp. 2389–2396, 2017. DOI: https://doi.org/10.1109/ACCESS.2017.2782884.
S. J. Winham, R. R. Freimuth, J. M. Biernacka. A weighted random forests approach to improve predictive performance. Statistical Analysis and Data Mining, vol. 6, no. 6, pp. 496–505, 2013. DOI: https://doi.org/10.1002/sam.11196.
Y. C. Chen, H. Ahn, J. J. Chen. High-dimensional canonical forest. Journal of Statistical Computation and Simulation, vol. 87, no. 5, pp. 845–854, 2017. DOI: https://doi.org/10.1080/00949655.2016.1231191.
H. F. Zhou, X. Z. Zhao, X. Wang. An effective ensemble pruning algorithm based on frequent patterns. Knowledge-Based Systems, vol. 56, pp. 79–85, 2014. DOI: https://doi.org/10.1016/j.knosys.2013.10.024.
Y. Zhang, S. Burer, W. N. Street. Ensemble pruning via semidefinite programming. Journal of Machine Learning Research, vol. 7, no. 1, pp. 1315–1338, 2006.
L. I. Kuncheva, J. J. Rodríguez. A weighted voting framework for classifiers ensembles. Knowledge and Information Systems, vol. 38, no. 2, pp. 259–275, 2014. DOI: https://doi.org/10.1007/s10115-012-0586-6.
A. Kabir, C. Ruiz, S. A. Alvarez. Mixed bagging: a novel ensemble learning framework for supervised classification based on instance hardness. In Proceedings of IEEE International Conference on Data Mining, IEEE, Singapore, Singapore, pp.1073–1078, 2018. DOI: https://doi.org/10.1109/ICDM.2018.00137.
L. V. Utkin, M. S. Kovalev, A. A. Meldo. A deep forest classifier with weights of class probability distribution subsets. Knowledge-based Systems, vol. 173, pp. 15–27, 2019. DOI: https://doi.org/10.1016/j.knosys.2019.02.022.
H. Reddy, N. Raj, M. Gala, A. Basava. Text-mining-based fake news detection using ensemble methods. International Journal of Automation and Computing, vol. 17, no. 2, pp. 210–221, 2020. DOI: https://doi.org/10.1007/s11633-019-1216-5.
W. G. Yi, J. Duan, M. Y. Lu. Double-layer Bayesian classifier ensembles based on frequent itemsets. International Journal of Automation and Computing, vol. 9, no. 2, pp. 215–220, 2012. DOI: https://doi.org/10.1007/s11633-012-0636-2.
G. Wang, J. X. Hao, J. Ma, H. B. Jiang. A comparative assessment of ensemble learning for credit scoring. Expert Systems with Applications, vol. 38, no. 1, pp. 223–230, 2011. DOI: https://doi.org/10.1016/j.eswa.2010.06.048.
F. Martínez-Plumed, R. B. Prudêncio, A. Martínez-Usó, J. Hernández-Orallo. Making sense of item response theory in machine learning. In Proceedings of the 22nd European Conference on Artificial Intelligence, IOS Press, The Hague, The Netherlands, pp. 1140–1148, 2016. DOI: https://doi.org/10.3233/978-1-61499-672-9-1140.
C. Zanon, C. S. Hutz, H. Yoo, R. K. Hambleton. An application of item response theory to psychological test development. Psicologia: Refflexão e Crítica, vol. 29, no. 1, Article number 18, 2016. DOI: https://doi.org/10.1186/s41155-016-0040-x.
H. L. Fu, G. Manogaran, K. Wu, M. Cao, S. Jiang, A. M. Yang. Intelligent decision-making of online shop** behavior based on internet of things. International Journal of Information Management, vol. 50, pp. 515–525, 2020. DOI: https://doi.org/10.1016/j.i**fomgt.2019.03.010.
W. R. Gilks, S. Richardson, D. J. Spiegelhalter. Markov Chain Monte Carlo in Practice. Boca Raton, USA: Chapman & Hall, CRC, 1995.
Y. Chen, T. S. Filho, R. B. C. Prudencio, T. Diethe, P. Flach. β3-IRT: a new item response model and its applications. ar**v: 1903.04016, 2019. https://arxiv.org/abs/1903.04016.
B. W. Junker, R. J. Patz, N. M. VanHoudnos. Markov chain Monte Carlo for item response models. Handbook of Item Response Theory, Volume Two: Statistical Tools, W. J. van der Linden, Ed., Boca Raton, USA: Chapman and Hall, CRC, pp. 271–325, 2016.
J. S. Kim, D. M. Bolt. Estimating item response theory models using Markov chain Monte Carlo methods. Educational Measurement: Issues and Practice, vol. 26, no. 4, pp. 38–51, 2007. DOI: https://doi.org/10.1111/j.1745-3992.2007.00107.x.
M. A. Tanner, W. H. Wong. The calculation of posterior distributions by data augmentation. Journal of the American Statistical Association, vol. 82, no. 398, pp. 528–540, 1987. DOI: https://doi.org/10.1080/01621459.1987.10478458.
J. H. Albert. Bayesian estimation of normal ogive item response curves using Gibbs sampling. Journal of Educational Statistics, vol. 17, no. 3, pp. 251–269, 1992. DOI: https://doi.org/10.3102/10769986017003251.
Y. Y. Sheng. Markov chain Monte Carlo estimation of normal ogive IRT models matlab. Journal of Statistical Software, vol. 25, no. 8, pp.1–15, 2008. DOI:https://doi.org/10.18637/jss.v025.i08.
Y. Y. Sheng. Bayesian estimation of the four-parameter IRT model using Gibbs sampling. International Journal of Quantitative Research in Education, vol. 2, no. 3–4, pp. 194–212, 2015. DOI: https://doi.org/10.1504/IJQRE.2015.071736.
Y. Noel, B. Dauvier. A beta item response model for continuous bounded responses. Applied Psychological Measurement, vol. 31, no. 1, pp. 47–73, 2007. DOI: https://doi.org/10.1177/0146621605287691.
J. C. Xu, Q. W. Ren, Z. Z. Shen. Prediction of the strength of concrete radiation shielding based on LS-SVM. Annals of Nuclear Energy, vol. 85, pp. 296–300, 2015. DOI: https://doi.org/10.1016/j.anucene.2015.05.030.
S. Borman. The expectation maximization algorithm: a short tutorial. Submmitted for Publication, vol. 41, 2004.
W. Deng, H. M. Zhao, L. Zou, G. Y. Li, X. H. Yang, D. Q. Wu. A novel collaborative optimization algorithm in solving complex optimization problems. Soft Computing, vol. 21, no. 15, pp. 4387–4398, 2017. DOI: https://doi.org/10.1007/s00500-016-2071-8.
M. H. Fang, X. H. Hu, T. T. He, Y. Wang, J. M. Zhao, X. J. Shen, J. Yuan. Prioritizing disease-causing genes based on network diffusion and rank concordance. In Proceedings of IEEE International Conference on Bioinformatics and Biomedicine, IEEE, Belfast, UK, pp. 242–247, 2014. DOI: https://doi.org/10.1109/BIBM.2014.6999162.
S. R. Safavian, D. Landgrebe. A survey of decision tree classifier methodology. IEEE Transactions on Systems, Man, and Cybernetics, vol. 21, no. 3, pp. 660–674, 1991. DOI: https://doi.org/10.1109/21.97458.
A. Liaw, M. Wiener. Classification and regression by randomforest. R News, vol. 2–3, pp. 18–22, 2002.
J. H. Friedman. Stochastic gradient boosting. Computational Statistics & Data Analysis, vol. 38, no. 4, pp. 367–378, 2002. DOI: https://doi.org/10.1016/S0167-9473(01)00065-2.
S. Mika, G. Ratsch, J. Weston, B. Scholkopf, K. R. Mullers. Fisher discriminant analysis with kernels. In Proceedings of IEEE Signal Processing Society Workshop, IEEE, Madison, USA, pp. 41–48, 1999. DOI: https://doi.org/10.1109/NNSP.1999.788121.
J. A. K. Suykens, J. Vandewalle. Least squares support vector machine classifiers. Neural Processing Letters, vol. 9, no. 3, pp. 293–300, 1999. DOI: https://doi.org/10.1023/A:1018628609742.
E. Bauer, R. Kohavi. An empirical comparison of voting classification algorithms: bagging, boosting, and variants. Machine Learning, vol. 36, no. 1–2, pp. 105–139, 1999. DOI: https://doi.org/10.1023/A:1007515423169.
H. Li, F. D. Chen, K. W. Cheng, Z. Z. Zhao, D. Z. Yang. Prediction of zeta potential of decomposed peat via machine learning: comparative study of support vector machine and artificial neural networks. International Journal of Electrochemical Science, vol. 10, no. 8, pp. 6044–6056, 2015.
Y. C. Chen, H. Ha, H. Kim, H. Ahn. Canonical forest. Computational Statistics, vol. 29, no. 3–4, pp. 849–867, 2014. DOI: https://doi.org/10.1007/s00180-013-0466-x.
Author information
Authors and Affiliations
Corresponding author
Additional information
Recommended by Associate Editor Matjaz Gams
Ziheng Chen received the B. Sc. degree in statistics from Renmin University of China, China in 2016. He is currently a Ph. D. degree candidate in Department of Applied Mathematics and Statistics, Stony Brook University, USA.
His research interests include reinforcement learning, recommending system, tree structure model and ensemble learning theory.
Hongshik Ahn received the B. Sc. degree in mathematics from Seoul National University, South Korea, and the Ph. D. degree in statistics from University of Wisconsin-Madison, USA in 1992. From 1992 to 1996, he was a mathematical statistician at the National Center for Toxicological Research, U.S. Food and Drug Administration, and a faculty member in the Department of Applied Mathematics and Statistics at Stony Brook University, USA from 1996 to present. He was the first Vice President of SUNY Korea for two years from 2012. Currently, he is a professor at Stony Brook University. He has published 2 books, 3 book chapters, over 70 papers in peer-reviewed journals, and 25 conference papers.
His research interests include classification of high-dimensional data, tree-structured regression modeling, survival analysis, and multi-step batch testing for infectious diseases.
Rights and permissions
About this article
Cite this article
Chen, Z., Ahn, H. Item Response Theory Based Ensemble in Machine Learning. Int. J. Autom. Comput. 17, 621–636 (2020). https://doi.org/10.1007/s11633-020-1239-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11633-020-1239-y