Artificial Intelligence Review

, Volume 33, Issue 1–2, pp 1–39 | Cite as

Ensemble-based classifiers

Article

Abstract

The idea of ensemble methodology is to build a predictive model by integrating multiple models. It is well-known that ensemble methods can be used for improving prediction performance. Researchers from various disciplines such as statistics and AI considered the use of ensemble methodology. This paper, review existing ensemble techniques and can be served as a tutorial for practitioners who are interested in building ensemble based systems.

Keywords

Ensemble of classifiers Supervised learning Classification Boosting 

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

References

  1. Arbel R, Rokach L (2006) Classifier evaluation under limited resources. Pattern Recognit Lett 27(14): 1619–1631CrossRefGoogle Scholar
  2. Banfield RE, Hall LO, Bowyer KW, Kegelmeyer WP (2007) A comparison of decision tree ensemble creation techniques. IEEE Trans Pattern Anal Mach Intell 29(1): 173–180CrossRefGoogle Scholar
  3. Bauer E, Kohavi R (1999) An empirical comparison of voting classification algorithms: bagging, boosting, and variants. Mach Learn 35: 1–38Google Scholar
  4. Bay S (1999) Nearest neighbor classification from multiple feature subsets. Intell Data Anal 3(3): 191–209CrossRefGoogle Scholar
  5. Biermann AW, Faireld J, Beres T (1982) Signature table systems and learning. IEEE Trans Syst Man Cybern 12(5): 635–648MATHCrossRefGoogle Scholar
  6. Breiman L (1996) Bagging predictors. Mach Learn 24(2): 123–140MATHMathSciNetGoogle Scholar
  7. Breiman L (1998) Arcing classifiers. Ann Stat 26(3): 801–849MATHCrossRefMathSciNetGoogle Scholar
  8. Breiman L (1999) Pasting small votes for classification in large databases and on-line. Mach Learn 36(2): 85–103CrossRefGoogle Scholar
  9. Breiman L (2001) Random forests. Mach Learn 45: 5–32MATHCrossRefGoogle Scholar
  10. Brodley CE (1995) Recursive automatic bias selection for classifier construction. Mach Learn 20: 63–94Google Scholar
  11. Bryll R, Gutierrez-Osuna R, Quek F (2003) Bagging: improving accuracy of classifier ensembles by using random feature subsets. Pattern Recognit 36: 1291–1302MATHCrossRefGoogle Scholar
  12. Brown G, Wyatt JL (2003) Negative correlation learning and the ambiguity family of ensemble methods. Proceedings of 4th international workshop, Mult Classifier Syst 2003, Guilford, UK, June 11–13, 2003, Lecture Notes in Computer Science, vol 2709, pp 266–275Google Scholar
  13. Brown G, Wyatt J, Harris R, Yao X (2005) Diversity creation methods: a survey and categorisation. Inf Fusion 6(1): 5–20CrossRefGoogle Scholar
  14. Buchanan BG, Shortliffe EH (1984) Rule based expert systems. Addison-Wesley, Reading 272–292Google Scholar
  15. Buntine W (1990) A theory of learning classification rules. Doctoral Dissertation. School of Computing Science, University of Technology. Sydney. AustraliaGoogle Scholar
  16. Buntine W (1996) Graphical models for discovering knowledge. In: Fayyad U, Piatetsky-Shapiro G, Smyth P, Uthurusamy R (eds) Advances in knowledge discovery and data mining. AAAI/MIT Press, Cambridge, pp 59–82Google Scholar
  17. Caruana R, Niculescu-Mizil A, Crew G, Ksikes A (2004) Ensemble selection from libraries of models, twenty-first international conference on Machine learning, July 04–08, 2004, Banff, Alberta, CanadaGoogle Scholar
  18. Chan PK, Stolfo SJ (1993) Toward parallel and distributed learning by meta-learning. In: AAAI Workshop in Knowledge Discovery in Databases, pp 227–240Google Scholar
  19. Chan PK, Stolfo SJ (1995) A comparative evaluation of voting and meta-learning on partitioned data, Proceeding of 12th international conference On machine learning ICML-95Google Scholar
  20. Chan PK, Stolfo SJ (1997) On the accuracy of meta-learning for scalable data mining. J Intell Inf Syst 8: 5–28CrossRefGoogle Scholar
  21. Charnes A, Cooper WW, Rhodes E (1978) Measuring the efficiency of decision making units. Eur J Oper Res 2(6): 429–444MATHCrossRefMathSciNetGoogle Scholar
  22. Chawla NV, Hall LO, Bowyer KW, Kegelmeyer WP (2004) Learning ensembles from bites: a scalable and accurate approach. J Mach Learn Res Arch 5: 421–451MathSciNetGoogle Scholar
  23. Chen K, Wang L, Chi H (1997) Methods of combining multiple classifiers with different features and their applications to text-independent speaker identification. Intern J Pattern Recognit Artif Intell 11(3): 417–445CrossRefGoogle Scholar
  24. Cherkauer KJ (1996) Human expert-level performance on a scientific image analysis task by a system using combined artificial neural networks. In: Working notes, integrating multiple learned models for improving and scaling machine learning algorithms workshop, thirteenth national conference on artificial intelligence. AAAI Press, Portland, ORGoogle Scholar
  25. Clark P, Boswell R (1991) Rule induction with CN2: some recent improvements. In: Proceedings of the European working session on learning, Pitman, pp 151–163Google Scholar
  26. Cohen S, Rokach L, Maimon O (2007) Decision tree instance space decomposition with grouped gain-ratio. Inf Sci 177: 3592–3612CrossRefGoogle Scholar
  27. Dasarathy BV, Sheela BV (1979) Composite classifier system design: concepts and methodology. Proc IEEE 67(5): 708–713CrossRefGoogle Scholar
  28. Derbeko P, El-Yaniv R, Meir R (2002) Variance optimized bagging. Eur Conf Mach LearnGoogle Scholar
  29. Džeroski S, Ženko B (2004) Is combining classifiers with stacking better than selecting the best one?. Mach Learn 54(3): 255–273MATHCrossRefGoogle Scholar
  30. Dietterich TG, Bakiri G (1995) Solving multiclass learning problems via error-correcting output codes. J Artif Intell Res 2: 263–286MATHGoogle Scholar
  31. Dimitriadou E, Weingessel A, Hornik K (2003) A cluster ensembles framework, Design and application of hybrid intelligent systems. IOS Press, AmsterdamGoogle Scholar
  32. Domingos P (1996) Using partitioning to speed up specific-to-general rule induction. In: Proceedings of the AAAI-96 workshop on integrating multiple learned models. AAAI Press, Cambridge, pp 29–34Google Scholar
  33. Frelicot C, Mascarilla L (2001) Reject strategies driver combination of pattern classifiersGoogle Scholar
  34. Freund S (1995) Boosting a weak learning algorithm by majority. Inf Comput 121(2): 256–285MATHCrossRefMathSciNetGoogle Scholar
  35. Freund Y, Schapire RE (1996) Experiments with a new boosting algorithm. In: Machine learning: proceedings of the thirteenth international conference, pp 325–332Google Scholar
  36. Friedman J, Hastie T, Tibshirani R (1998) Additive logistic regression: a statistical view of boostingGoogle Scholar
  37. Gama J (2004) A linear-bayes classifier. In: Monard C (ed) Advances on artificial intelligence—SBIA2000. LNAI 1952, Springer, Berlin, pp 269–279Google Scholar
  38. Gams M (1989) New measurements highlight the importance of redundant knowledge. In: European working session on learning, Montpeiller, France, PitmanGoogle Scholar
  39. Garcia-Pddrajas N, Garcia-Osorio C, Fyfe C (2007) Nonlinear boosting projections for ensemble construction. J Mach Learn Res 8: 1–33MathSciNetGoogle Scholar
  40. Hampshire JB, Waibel A (1992) The meta-Pi network—building distributed knowledge representations for robust multisource pattern-recognition. Pattern Anal Mach Intell 14(7): 751–769CrossRefGoogle Scholar
  41. Hansen LK, Salamon P (1990) Neural network ensembles. IEEE Trans Pattern Anal Mach Intell 12(10): 993–1001CrossRefGoogle Scholar
  42. Hansen J (2000) Combining predictors. Meta machine learning methods and bias, variance & ambiguity decompositions. PhD Dissertation. Aurhus UniversityGoogle Scholar
  43. Ho TK (1998) The random subspace method for constructing decision forests. IEEE Trans Pattern Anal Mach Intell 20(8): 832–844CrossRefGoogle Scholar
  44. Holmstrom L, Koistinen P, Laaksonen J, Oja E (1997) Neural and statistical classifiers—taxonomy and a case study. IEEE Trans Neural Netw 8: 5–17CrossRefGoogle Scholar
  45. Hsu CW, Lin CJ (2002) A comparison of methods for multi-class support vector machines. IEEE Trans Neural Netw 13(2): 415–425CrossRefGoogle Scholar
  46. Hu Q, Yu D, Xie Z, Li X (2007) EROS: ensemble rough subspaces. Pattern Recognit 40: 3728–3739MATHCrossRefGoogle Scholar
  47. Hu X (2001) Using rough sets theory and database operations to construct a good ensemble of classifiers for data mining applications. ICDM01. pp 233–240Google Scholar
  48. Islam MM, Yao X, Murase K (2003) A constructive algorithm for training cooperative neural network ensembles. IEEE Trans Neural Netw 14(4): 820–834CrossRefGoogle Scholar
  49. Jenkins R, Yuhas BPA (1993) Simplified neural network solution through problem decomposition: The case of Truck backer-upper. IEEE Trans Neural Netw 4(4): 718–722CrossRefGoogle Scholar
  50. Johansen TA, Foss BA (1992) A narmax model representation for adaptive control based on local model—Modeling. Identif Control 13(1): 25–39Google Scholar
  51. Jordan MI, Jacobs RA (1994) Hierarchical mixtures of experts and the EM algorithm. Neural Comput 6: 181–214CrossRefGoogle Scholar
  52. Kamath C, Cantu-Paz E (2001) Creating ensembles of decision trees through sampling, Proceedings, 33rd symposium on the interface of computing science and statistics, Costa Mesa, CA, June 2001Google Scholar
  53. Kamath C, Cantú E, Littau D (2002) Approximate splitting for ensembles of trees using histograms. In: Second SIAM international conference on data mining (SDM-2002)Google Scholar
  54. Kohavi R (1996) Scaling up the accuracy of naive-bayes classifiers: a decision-tree hybrid. In: Proceedings of the second international conference on knowledge discovery and data mining, pp 114–119Google Scholar
  55. Kolen JF, Pollack JB (1991) Back propagation is sesitive to initial conditions. In: Advances in neural information processing systems, vol 3. Morgan Kaufmann, San Francisco, pp 860–867Google Scholar
  56. Krogh A, Vedelsby J (1995) Neural network ensembles, cross validation and active learning. Adv Neural Inf Process Syst 7: 231–238Google Scholar
  57. Kuncheva L (2005) Combining pattern classifiers. Wiley Press, New YorkGoogle Scholar
  58. Kuncheva L, Whitaker C (2003) Measures of diversity in classifier ensembles and their relationship with ensemble accuracy. Mach Learn 51(2):181–207MATHCrossRefGoogle Scholar
  59. Kuncheva LI (2005) Diversity in multiple classifier systems (Editorial). Inf Fusion 6(1): 3–4CrossRefMathSciNetGoogle Scholar
  60. Kusiak A (2000) Decomposition in data mining: an industrial case study. IEEE Trans Electron Packaging Manuf 23(4): 345–353CrossRefGoogle Scholar
  61. Langdon WB, Barrett SJ, Buxton BF, (2002) Combining decision trees and neural networks for drug discovery. In: Genetic programming, proceedings of the 5th European conference, EuroGP 2002, Kinsale, Ireland, pp 60–70Google Scholar
  62. Liu Y (2005) Generate different neural networks by negative correlation learning. ICNC 1: 149–156Google Scholar
  63. Liu H, Mandvikar A, Mody J (2004) An empirical study of building compact ensembles. WAIM 622–627Google Scholar
  64. Long C (2003) Bi-decomposition of function sets using multi-valued logic, Eng Doc Dissertation, Technischen Universitat Bergakademie FreibergGoogle Scholar
  65. Maimon O, Rokach L (2005) Decomposition methodology for knowledge discovery and data mining: theory and applications. World Scientific, SingaporeMATHGoogle Scholar
  66. Maimon O, Rokach L (2002) Improving supervised learning by feature decomposition. In: Proceedings of foundations of information and knowledge systems, Salzan Castle, Germany, pp 178–196Google Scholar
  67. Margineantu D, Dietterich T (1997) Pruning adaptive boosting. In: Proceedings of fourteenth international conference machine learning, pp 211–218Google Scholar
  68. Melville P, Mooney RJ (2003) Constructing diverse classifier ensembles using artificial training examples. IJCAI 505–512Google Scholar
  69. Merler S, Caprile B, Furlanello C (2007) Parallelizing AdaBoost by weights dynamics. Comput Stat Data Anal 51: 2487–2498MATHCrossRefMathSciNetGoogle Scholar
  70. Merz CJ (1999) Using correspondence analysis to combine classifier. Mach Learn 36(1–2): 33–58CrossRefGoogle Scholar
  71. Michalski RS, Tecuci G (1994) Machine learning, a multistrategy approach. Morgan Kaufmann, San FranciscoGoogle Scholar
  72. Michie D (1995) Problem decomposition and the learning of skills, In: Proceedings of the European conference on machine learning, Springer, Berlin, pp 17–31Google Scholar
  73. Mitchell T (1980) The need for biases in learning generalizations. Technical Report CBM-TR-117, Rutgers University, Department of Computer Science, New BrunswickGoogle Scholar
  74. Nowlan SJ, Hinton GE (1991) Evaluation of adaptive mixtures of competing experts. In: Lippmann RP, Moody JE, Touretzky DS (eds) Advances in neural information processing systems, vol 3. Morgan Kaufmann Publishers, San Francisco, pp 774–780Google Scholar
  75. Ohno-Machado L, Musen MA (1997) Modular neural networks for medical prognosis: quantifying the benefits of combining neural networks for survival prediction. Connect Sci 9(1): 71–86CrossRefGoogle Scholar
  76. Opitz D (1999) Feature selection for ensembles. In: Proceedings of 16th National Conference on Artificial Intelligence, AAAI. pp 379–384Google Scholar
  77. Opitz D, Shavlik J (1996) Generating accurate and diverse members of a neuralá1network ensemble. In: Touretzky DS, Mozer MC, Hasselmo ME (eds) Advances in neural information processing systems, vol 8. The MIT Press, Cambridge, pp 535–541Google Scholar
  78. Parmanto B, Munro PW, Doyle HR (1996) Improving committee diagnosis with resampling techinques. In: Touretzky DS, Mozer MC, Hesselmo ME (eds) Advances in neural information processing systems, vol 8. MIT Press, Cambridge, pp 882–888Google Scholar
  79. Partridge D, Yates WB (1996) Engineering multiversion neural-net systems. Neural Comput 8(4): 869–893CrossRefGoogle Scholar
  80. Passerini A, Pontil M, Frasconi P (2004) New results on error correcting output codes of kernel machines. IEEE Trans Neural Netw 15: 45–54CrossRefGoogle Scholar
  81. Peng F, Jacobs RA, Tanner MA (1995) Bayesian inference in mixtures-of-experts and hierarchical mixturesof-experts models with an application to speech recognition. J Am Stat Assoc 91(435):953–960CrossRefGoogle Scholar
  82. Polikar R (2006) Ensemble based systems in decision making. IEEE Circuits Syst Mag 6(3): 21–45CrossRefGoogle Scholar
  83. Prodromidis AL, Stolfo SJ, Chan PK (1999) Effective and efficient pruning of metaclassifiers in a distributed Data Mining system. Technical report CUCS-017-99, Columbia UniversityGoogle Scholar
  84. Prodromidis AL, Stolfo SJ (2001) Cost complexity-based pruning of ensemble classifiers. Knowl Inf Syst 3(4): 449–469MATHCrossRefGoogle Scholar
  85. Provost FJ, Kolluri V (1997) A survey of methods for scaling up inductive learning algorithms. In: Proceeding of 3rd international conference on knowledge discovery and data miningGoogle Scholar
  86. Quinlan JR (1993) C4.5: programs for machine learning. Morgan Kaufmann, Los AltosGoogle Scholar
  87. Quinlan JR (1996) Bagging, Boosting, and C4.5. In: Proceedings of the thirteenth national conference on artificial intelligence, pp 725–730Google Scholar
  88. Rahman AFR, Fairhurst MC (1997) A new hybrid approach in combining multiple experts to recognize handwritten numerals. Pattern Recognit Lett 18: 781–790CrossRefGoogle Scholar
  89. Ramamurti V, Ghosh J (1999) Structurally adaptive modular networks for non-stationary environments. IEEE Trans Neural Netw 10(1): 152–160CrossRefGoogle Scholar
  90. Rokach L (2006) Decomposition methodology for classification tasks—A meta decomposer framework. Pattern Anal Appl 9: 257–271CrossRefMathSciNetGoogle Scholar
  91. Rokach L (2008) Genetic algorithm-based feature set partitioning for classification problems. Pattern Recognit 41(5): 1676–1700MATHCrossRefGoogle Scholar
  92. Rokach L (2009) Collective-agreement-based pruning of ensembles. Comput Stat Data Anal 53(4): 1015–1026CrossRefMATHGoogle Scholar
  93. Rokach L, Arbel R, Maimon O (2006) Selective voting—getting more for less in sensor fusion. Intern J Pattern Recognit Artif Intell 20(3): 329–350CrossRefGoogle Scholar
  94. Rokach L, Maimon O (2005a) Top down induction of decision trees classifiers: A survey. IEEE SMC Trans Part C 35(4):476–487CrossRefGoogle Scholar
  95. Rokach L, Maimon O (2005b) Feature Set decomposition for decision trees. J Intell Data Anal 9(2): 131–158Google Scholar
  96. Rokach L, Maimon O (2008) Data mining with decision trees: theory and applications. World Scientific Publishing, SingaporeMATHGoogle Scholar
  97. Rokach L, Maimon O, Arad O (2005) Improving supervised learning by sample decomposition. Int J Comput Intell Appl 5(1): 37–54CrossRefGoogle Scholar
  98. Rokach L, Maimon O, Lavi I (2003) Space decomposition in data mining: A clustering approach. In: Proceedings of the 14th international symposium on methodologies for intelligent systems, Maebashi, Japan, Lecture Notes in Computer Science, Springer, Berlin, pp 24–31Google Scholar
  99. Rudin C, Daubechies I, Schapire RE (2004) The dynamics of Adaboost: cyclic behavior and convergence of margins. J Mach Learn Res 5: 1557–1595MathSciNetGoogle Scholar
  100. Rosen BE (1996) Ensemble learning using decorrelated neural networks. Connect Sci 8(3): 373–384CrossRefGoogle Scholar
  101. Samuel A (1967) Some studies in machine learning using the game of checkers II: Recent progress. IBM J Res Develop 11: 601–617CrossRefGoogle Scholar
  102. Saaty X (1996) The analytic hierarchy process: A 1993 overview. Cent Eur J Oper Res Econ 2(2): 119–137Google Scholar
  103. Schaffer C (1993) Selecting a classification method by cross-validation. Mach Learn 13(1): 135–143Google Scholar
  104. Schapire RE (1990) The strength of weak learnability. Mach Learn 5(2): 197–227Google Scholar
  105. Seewald AK, Fürnkranz J (2001) Grading classifiers, Austrian research institute for Artificial intelligenceGoogle Scholar
  106. Sharkey A (1996) On combining artificial neural nets. Connect Sci 8: 299–313CrossRefGoogle Scholar
  107. Sharkey N, Neary J, Sharkey A (1995) Searching weight space for backpropagation solution types, current trends in connectionism: Proceedings of the 1995 Swedish conference on connectionism, pp 103–120Google Scholar
  108. Shilen S (1990) Multiple binary tree classifiers. Pattern Recognit 23(7): 757–763CrossRefGoogle Scholar
  109. Skurichina M, Duin RPW (2002) Bagging, boosting and the random subspace method for linear classifiers. Pattern Anal Appl 5(2): 121–135MATHCrossRefMathSciNetGoogle Scholar
  110. Sohn SY, Choi H (2001) Ensemble based on data envelopment analysis. ECML Meta Learning workshopGoogle Scholar
  111. Tamon C, Xiang J (2000) On the boosting pruning problem. In: Proceedings of the 11th European conference on machine learning, pp 404–412Google Scholar
  112. Towell G, Shavlik J (1994) Knowledge-based artificial neural networks. Artif Intell 70: 119–165MATHCrossRefGoogle Scholar
  113. Tsoumakas G, Partalas I, Vlahavas I (2008) A taxonomy and short review of ensemble selection. In: ECAI 2008, workshop on supervised and unsupervised ensemble methods and their applicationsGoogle Scholar
  114. Tsymbal A, Puuronen S (2002) Ensemble feature selection with the simple bayesian classification in medical diagnostics, In: Proceedings of 15th IEEE symposium on Computer-Based Medical Systems CBMS2002, IEEE CS Press, Maribor, Slovenia, pp 225–230Google Scholar
  115. Tukey JW (1977) Exploratory data analysis. Addison-Wesley, ReadingMATHGoogle Scholar
  116. Tumer K, Ghosh J (1996) Error correlation and error reduction in ensemble classifiers. Connection science, special issue on combining artificial neural networks: ensemble approaches. 8(3–4): 385–404Google Scholar
  117. Tumer K, Ghosh J (2000) Robust Order Statistics based Ensembles for Distributed Data Mining. In: Kargupta H, Chan P (eds) Advances in distributed and parallel knowledge discovery. AAAI/MIT Press, Cambridge, pp 185–210Google Scholar
  118. Tumer K, Oza CN (2003) Input decimated ensembles. Pattern Anal Appl 6: 65–77MATHCrossRefMathSciNetGoogle Scholar
  119. Wang W, Jones P, Partridge D (2000) Diversity between neural networks and decision trees for building multiple classifier systems, In: Proceeding of international workshop on multiple classifier systems (LNCS 1857), Springer, Calgiari, pp 240–249Google Scholar
  120. Weigend AS, Mangeas M, Srivastava AN (1995) Nonlinear gated experts for time-series—discovering regimes and avoiding overfitting. Int J Neural Syst 6(5): 373–399CrossRefGoogle Scholar
  121. Weston J, Watkins C (1999) Support vector machines for multi-class pattern recognition. In: Verleysen M (ed) Proceedings of the 7th European symposium on artificial neural networks (ESANN-99), Bruges, Belgium, pp 219–224Google Scholar
  122. Windeatt T, Ardeshir G (2001) An Empirical comparison of pruning methods for ensemble classifiers, IDA2001, LNCS 2189, pp 208–217Google Scholar
  123. Woods K, Kegelmeyer W, Bowyer K (1997) Combination of multiple classifiers using local accuracy estimates. IEEE Trans Pattern Anal Mach Intell 19: 405–410CrossRefGoogle Scholar
  124. Wolpert DH (1992) Stacked generalization. In: (eds) Neural Networks, vol 5. Pergamon Press, Oxford, pp 241–259Google Scholar
  125. Yanim S, Kamel MS, Wong AKC, Wang Y (2007) Cost-sensitive boosting for classification of imbalanced data. Pattern Recognition (40): 3358–3378Google Scholar
  126. Yates W, Partridge D (1996) Use of methodological diversity to improve neural network generalization. Neural Comput Appl 4(2): 114–128CrossRefGoogle Scholar
  127. Zhang Y, Burer S, Street WN (2006) Ensemble pruning via semi-definite programming. J Mach Learn Res 7: 1315–1338MathSciNetGoogle Scholar
  128. Zhang CX, Zhang JS (2008) A local boosting algorithm for solving classification problems. Comput Stat Data Anal 52(4): 1928–1941MATHCrossRefGoogle Scholar
  129. Zhou ZH, Tang W (2003) Selective ensemble of decision trees. In: Wang G, Liu Q, Yao Y, Skowron A (eds) Rough sets, fuzzy sets, data mining, and granular computing, 9th international conference, RSFDGrC, Chongqing, China, Proceedings. Lecture Notes in Computer Science 2639, pp 476–483Google Scholar
  130. Zhou ZH, Wu J, Tang W (2002) Ensembling neural networks: many could be better than all. Artif Intell 137: 239–263MATHCrossRefMathSciNetGoogle Scholar
  131. Zupan B, Bohanec M, Demsar J, Bratko I (1998) Feature transformation by function decomposition. IEEE Intell Syst Appl 13: 38–43CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media B.V. 2009

Authors and Affiliations

  1. 1.Department of Information System EngineeringBen-Gurion University of the NegevBeer-ShevaIsrael

Personalised recommendations