Traditional Two-way classification algorithms fail to represent the uncertainty in data, whereas Three-way classification algorithms represent uncertainty regarding the boundary region. In Two-way classification algorithms, based on the specific strategy of the model, objects are classified and the classification accuracy is improved by the appropriate selection of the model. The Three-way classification models extend to various methods based on rough sets, fuzzy sets and interval sets. Among these methods, Probabilistic rough sets work on the basis of equivalence classes which impart the indiscernibility among the objects. Adding a proper dimensionality for the object description reduces the uncertainty which leads to definite classifications. By focusing on the benefits of both types of algorithms, this paper proposes a hybridized method using Two-way classifications and Probabilistic rough sets to reduce the misclassification error and uncertainty. For that, proper Two-way classification is adapted to generate a Three-way classification, and the boundary region is re-examined to reduce the uncertainty by incorporating Information-theoretic rough sets. The experiment on the spam data sets proves that the proposed method increases the accuracy and coverage with minimum uncertainty compared to the existing methods. This work showcases the advantages of both Two-way as well as Three-way probabilistic rough set algorithms.
VapnikVGuyonIHastieT. Support vector machines. Mach Learn1995; 20: 273–297.
2.
RishI, et al. An empirical study of the naive bayes classifier. In: IJCAI 2001 workshop on empirical methods in artificial intelligence, volume 3. 2001, pp.41–46.
SafavianSRLandgrebeD. A survey of decision tree classifier methodology. IEEE Trans Syst Man Cybernet1991; 21: 660–674.
5.
LiYZhangLXuY, et al. Enhancing binary classification by modeling uncertain boundary in three-way decisions. IEEE Trans Knowl Data Eng2017; 29: 1438–1451.
6.
YaoJCornelisCWangG, et al. Uncertainty and three-way decision in data science. Int J Approx Reason2023; 162: 109024.
7.
JiaXShangL. Three-way decisions versus two-way decisions on filtering spam email. In: Transactions on rough sets XVIII. Springer, 2014. pp.69–91. DOI: 10.1007/978-3-662-44680-5_5.
8.
YaoY. Two semantic issues in a probabilistic rough set model. Fundam Inform2011; 108: 249–265.
9.
YaoYDengX. A granular computing paradigm for concept learning. In: Emerging paradigms in machine learning. Springer, 2013. pp.307–326. DOI: 10.1007/978-3-642-28699-5_12.
10.
DengX. Three-way classification models. Saskatchewan, Canada: The University of Regina, 2015.
11.
WongSLingrasP. A decision-theoretic rough set model. In: Proceedings of the 5th international symposium on methodologies for intelligent systems. North-Holland. 1990, pp.17–24.
12.
DengXYaoY. An information-theoretic interpretation of thresholds in probabilistic rough sets. In: International conference on rough sets and knowledge technology. Springer, 2012, pp.369–378. DOI: 10.1007/978-3-642-31900-6_46.
13.
DengXYaoY. A multifaceted analysis of probabilistic three-way decisions. Funda Inform2014; 132: 291–313.
YaoYZhouB. Naive bayesian rough sets. In: International conference on rough sets and knowledge technology. Springer, 2010, pp.719–726. DOI: 10.1007/978-3-642-16248-0_97.
16.
GrecoSMatarazzoBSłowińskiR. Parameterized rough set model using rough membership and bayesian confirmation measures. Int J Approx Reason2008; 49: 285–300.
17.
GrecoSMatarazzoBSłowińskiR. Rough membership and bayesian confirmation measures for parameterized rough sets. In: International workshop on rough sets, fuzzy sets, data mining, and granular-soft computing. Springer, 2005, pp.314–324. DOI: 10.1007/11548669_33.
18.
SlezakDZiarkoW. Bayesian rough set model. Proc of FDM2002; 131–135.
ZhangYYaoJ. Determining three-way decision regions with gini coefficients. In: International conference on rough sets and current trends in computing. Springer, 2014, pp.160–171. DOI: 10.1007/978-3-319-08644-6_17.
21.
ZhangYYaoJ. Gini objective functions for three-way classifications. Int J Approx Reason2017; 81: 103–114.
22.
GaoCYaoY. Determining thresholds in three-way decisions with chi-square statistic. In: International joint conference on rough sets. Springer, 2016, pp.272–281. DOI: 10.1007/978-3-319-47160-0_25.
23.
AzamNYaoJT. Variance based determination of three-way decisions using probabilistic rough sets. In: International joint conference on rough sets. Springer, 2016, pp.209–218. DOI: 10.1007/978-3-319-47160-0_19.
24.
AhmedNAminRAldabbasH, et al. Machine learning techniques for spam detection in email and IoT platforms: Analysis and research challenges. Sec Commun Networks2022; 2022: 1–19.
25.
CrawfordMKhoshgoftaarTMPrusaJD, et al. Survey of review spam detection using machine learning techniques. J Big Data2015; 2: 1–24.
26.
MadhavanMVPandeSUmekarP, et al. Comparative analysis of detection of email spam with the aid of machine learning approaches. In: IOP conference series: materials science and engineering, Vol. 1022. IOP Publishing, 2021, p.012113. DOI: 10.1088/1757-899X/1022/1/012113.
27.
ZhouBYaoYLuoJ. A three-way decision approach to email spam filtering. In: Canadian conference on artificial intelligence. Springer, 2010, pp.28–39. DOI: 10.1007/978-3-642-13059-5_6.
28.
JiaXZhengKLiW, et al. Three-way decisions solution to filter spam email: an empirical study. In: International conference on rough sets and current trends in computing. Springer, 2012, pp.287–296. DOI: 10.1007/978-3-642-32115-3_34.
LiJDengXYaoY. Multistage email spam filtering based on three-way decisions. In: International conference on rough sets and knowledge technology. Springer, 2013, pp.313–324. DOI: 10.1007/978-3-642-41299-8_30.
31.
ZhangYLiuPYaoJ. Three-way email spam filtering with game-theoretic rough sets. In: 2019 International conference on computing, networking and communications (ICNC). IEEE, 2019, pp.552–556. DOI: 10.1109/ICCNC.2019.8685642.
HasanMAMNasserMAhmadS, et al. Feature selection for intrusion detection using random forest. J Inform Sec2016; 7: 129–140.
35.
DuttaSGhatakSDeyR, et al. Attribute selection for improving spam classification in online social networks: a rough set theory-based approach. Soc Netw Anal Min2018; 8: 1–16.
36.
SainiMVermaSSharanA. Multi-view ensemble learning using rough set based feature ranking for opinion spam detection. In: Advances in computer communication and computational sciences: Proceedings of IC4S 2017, Vol. 1. Springer, 2019, pp.3–12. DOI: 10.1007/978-981-13-0341-8_1.
37.
SabuMKRajuG. Studies on rough set theory with applications to data mining. Kottayam, India: Shodhganga, MG University, 2013.
38.
IshibuchiHYamamotoTNakashimaT. Fuzzy data mining: Effect of fuzzy discretization. In: Proceedings 2001 IEEE international conference on data mining. IEEE, 2001, pp.241–248. DOI: 10.1109/ICDM.2001.989525.
39.
ChoiYSMoonBR. Feature selection in genetic fuzzy discretization for the pattern classification problems. IEICE Trans Inform Syst2007; 90: 1047–1054.
40.
PawlakZ. Rough sets. Int J Comput Inform Sci1982; 11: 341–356.
41.
Ujwala RekhaJShahu ChatrapatiKVinaya BabuA. Game theory and its applications in machine learning. In: Information systems design and intelligent applications: Proceedings of third international conference INDIA 2016, Vol. 3. Springer, 2016, pp.195–207. DOI: 10.1007/978-81-322-2757-1_21.
42.
HabibMAMohS. Game theory-based routing for wireless sensor networks: a comparative survey. Appl Sci2019; 9: 2896.
43.
DoCTTranNHHongC, et al. Game theory for cyber security and privacy. ACM Comput Surv (CSUR)2017; 50: 1–37.
44.
ChiCWangYTongX, et al. Game theory in internet of things: a survey. IEEE Int Things J2021; 9: 12125–12146.
45.
YanRPengLXieY, et al. Rough set-game theory information mining model considering opponents’ information. Electronics2022; 11: 244. DOI: https://doi.org/10.3390/electronics11020244.
46.
BashirZMahnazSAbbas MalikMG. Conflict resolution using game theory and rough sets. Int J Intell Syst2021; 36: 237–259.
47.
AzamNZhangYYaoJ. Evaluation functions and decision conditions of three-way decisions with game-theoretic rough sets. Eur J Oper Res2017; 261: 704–714.
48.
LiuP. Incorporating three-way email spam filtering with game-theoretic rough sets. Saskatchewan, Canada: The University of Regina, 2019.
49.
ZhouYZhangYYaoJ. Satirical news detection with semantic feature extraction and game-theoretic rough sets. In: Foundations of intelligent systems: 25th international symposium, ISMIS 2020, Graz, Austria, September 23–25, 2020, Proceedings. Springer, 2020, pp.123–135. DOI: 10.1007/978-3-030-59491-6_12.
50.
AzamNYaoJ. Analyzing uncertainties of probabilistic rough set regions with game-theoretic rough sets. Int J Approx Reason2014; 55: 142–155.
51.
MaldonadoSPetersGWeberR. Credit scoring using three-way decisions with probabilistic rough sets. Inf Sci2020; 507: 700–714.
52.
Novo-LouresMRuano-OrdasDPavonR, et al. Enhancing representation in the context of multiple-channel spam filtering. Inf Process Manag2022; 59: 102812.
53.
XuJZhangYMiaoD. Three-way confusion matrix for classification: a measure driven view. Inf Sci2020; 507: 772–794.
54.
FriedmanM. The use of ranks to avoid the assumption of normality implicit in the analysis of variance. J Am Stat Assoc1937; 32: 675–701.
55.
FriedmanM. A comparison of alternative tests of significance for the problem of m rankings. Ann Math Stat1940; 11: 86–92.
56.
ConoverW. Practical nonparametric statistics. New York, NY: John Wiley & Sons, Inc, 1999.
57.
HuM. Sequential approaches to three-way data analytics. Saskatchewan, Canada: The University of Regina, 2019.
58.
MohananGPadmanabhanDMAnishaGS. Analyzing random forest, naive bayes, and svm to filter spam emails across multiple datasets. In: Kumar A and Mozar S (eds.) ICCCE 2021. Singapore: Springer Nature Singapore. ISBN 978-981-16-7985-8, 2022, pp.325–332. DOI: 10.1007/978-981-16-7985-8_34.
59.
VinithaVSRenukaDK. Feature selection techniques for email spam classification: A survey. In: Kumar LA, Jayashree LS and Manimegalai R, (eds.) Proceedings of international conference on artificial intelligence, smart grid and smart city applications. Cham: Springer International Publishing, 2020, pp.925–935. DOI: 10.1007/978-3-030-24051-6_86.
60.
IranzadRLiuX. A review of random forest-based feature selection methods for data science education and applications. Int J Data Sci Anal2024; 2: 927312.
61.
DedeturkBKAkayB. Spam filtering using a logistic regression model trained by an artificial bee colony algorithm. Appl Soft Comput2020; 91: 106229.
62.
BassiouniMAliMEl-DahshanE. Ham and spam e-mails classification using machine learning techniques. J Appl Sec Res2018; 13: 315–331.
63.
SumathiVPVanithaVKalaiselviR, et al. Performance comparison of machine learning algorithms in short message service spam classification. In: Proceedings of the 2023 international conference on machine learning and data engineering.2023, pp.45–50. DOI: https://doi.org/10.1109/ICMLDE.2023.00010.
64.
AwadWAELseuofiS. Machine learning methods for spam e-mail classification. AIRCC’s Int J Comput Sci Inf Technol2011; 3: 173–184.
65.
MajeedBAzamNYaoJT. Thresholds determination for probabilistic rough sets with genetic algorithms. In: Rough sets and knowledge technology: 9th international conference, RSKT 2014, Shanghai, China, October 24-26, 2014, Proceedings 9. Springer, 2014, pp.693–704. DOI: 10.1007/978-3-319-11740-9_64.
66.
SoumyaTSabuM. Optimisation of thresholds in probabilistic rough sets with artificial bee colony algorithm. Fuzzy Inf Eng2021; 13: 522–539.
