The most important asset of semi-supervised learning (SSL) methods is the use of available unlabeled data combined with an enough smaller set of labeled examples, so as to increase the classification accuracy compared with the default procedure of supervised methods, in which during the training only the labeled data are used. The encapsulation of classifier ensembles that produce different models through training process into semi-supervised schemes seems to be a promising strategy for enhanced learning ability. In this work, a Self-trained Rotation Forest (Self-RotF) algorithm and a variant of this (Weighted-Self-RotF) are presented. We performed an in depth comparison with other well-known semi-supervised classification methods on standard benchmark datasets and after having tested their performance with statistical tests, we finally reached to the point that the presented technique had better accuracy in most cases.
ZhuX. and GoldbergA.B., Introduction to semi-supervised learning, Morgan & Claypool3(1) (2009).
2.
ZhouZ.-H. and LiM., Tri-training: Exploiting unlabeled data using three classifiers, IEEE Trans Knowl Data Eng17(11) (2005), 1529–1541.
3.
MahdhaouiA. and ChetouaniM., Supervised and semi-supervised infant-directed speech classification for parent-infant interaction analysis, Speech Commun53(9–10) (2011), 1149–1161.
4.
RosenbergC., HebertM. and SchneidermanH., Semi-Supervised Self-Training of Object Detection Models, in 2005 Seventh IEEE Workshops on Applications of Computer Vision - Volume 1, vol. 1, 2005, pp. 29–36.
5.
Hassan KhanF.,
QamarU. and BashirS., Building Normalized SentiMI to enhance semi-supervised sentiment analysis, J Intell Fuzzy Syst29(5) (2015), 1805–1816.
6.
ÖzçiftA. and
GultenA., Classifier ensemble construction with rotation forest to improve medical diagnosis performance of machine learning algorithms, Comput Methods Programs Biomed104(3) (2011), 443–451.
7.
YuenP.C., A boosted Co-training algorithm for human action recognition, IEEE Trans Circuits Syst Video Technol21(9) (2011), 1203–1213.
8.
HuangT., YuY., GuoG. and LiK., A classification algorithm based on local cluster centers with a few labeled training examples, Knowledge-Based Syst23(6) (2010), 563–571.
9.
TrigueroI., GarcíaS. and HerreraF., Self-labeled techniques for semi-supervised learning: Taxonomy, software and empirical study, Knowl Inf Syst42(2) (2013), 245–284.
10.
SchwenkerF. and TrentinE., Pattern classification and clustering: A review of partially supervised learning approaches, Pattern Recognit Lett37(1) (2014), 4–14.
11.
TanhaJ. and vanM., Someren and H. Afsarmanesh, Semi-supervised self-training for decision tree classifiers, Int J Mach Learn Cybern (2015), 1–16.
12.
RodriguezJ.J., KunchevaL.I. and AlonsoC.J., Rotation forest: A New classifier ensemble method, IEEE Trans Pattern Anal Mach Intell28(10) (2006), 1619–1630.
13.
DuP., SamatA., WaskeB., LiuS. and LiZ., Random Forest and Rotation Forest for fully polarized SAR image classification using polarimetric and spatial features, ISPRS J Photogramm Remote Sens105 (2015), 38–53.
HanJ., KamberM., Data mining: Concepts and techniques. Morgan Kaufmann Publishers, 2001.
16.
KunchevaL.I., RodríguezJ.J., An Experimental Study on Rotation Forest Ensembles, in Multiple Classifier Systems, Berlin, Heidelberg: Springer Berlin Heidelberg, 2007, pp. 459–468.
17.
LiM. and ZhouZ.H., Improve computer-aided diagnosis with machine learning techniques using undiagnosed samples, IEEE Trans Syst Man, Cybern Part ASystems Humans37(6) (2007), 1088–1098.
18.
LiM. and ZhouZ., SETRED: Self-training with Editing, LNAI3518 (2005), 611–621.
19.
BlumA. and MitchellT., Combining labeled and unlabeled data with co-training, in Proceedings of the Eleventh Annual Conference on Computational Learning Theory - COLT’ 98, 1998, pp. 92–100.
20.
ChapelleO., SchölkopfB. and ZienA., Semi-supervised learning, Interdiscip Sci2(2) (2006), 151–155.
21.
TrigueroI., GarciaS. and HerreraF., SEG-SSC: A framework based on synthetic examples generation for self-labeled semi-supervised classification, IEEE Trans Cybern45(4) (2015), 622–634.
22.
SunS., A survey of multi-view machine learning, Neural Comput Appl23(7–8) (2013), 2031–2038.
23.
DidaciL., FumeraG. and RoliF., Analysis of co-training algorithm with very small training sets, Lect Notes Comput Sci (including Subser Lect Notes Artif Intell Lect Notes Bioinformatics), vol. 7626 LNCS, 2012, pp. 719–726.
24.
SunS. and ZhangQ., Multiple-view multiple-learner semi-supervised learning, Neural Process Lett34(3) (2011), 229–240.
25.
NigamK. and GhaniR., Analyzing the effectiveness and applicability of co-training, Proc Ninth Int Conf Inf Knowl Manag - CIKM ’00, 2000, pp. 86–93.
26.
SunS. and JinF., Robust Co-training, Int J Pattern Recognit Artif Intell25(07) (2011), 1113–1126.
27.
HadyM. and SchwenkerF., Co-training by committee: A generalized framework for semi-supervised learning with committees, Int J Softw Informatics2(2) (2008), 95–124.
28.
XuC., TaoD. and XuC., A survey on multi-view learning, Cvpr36(8) (2015), 300072.
29.
HotellingH., Relations between two sets of variates, Biometrika28(3–4) (1936), 321–377.
30.
ZhouY. and GoldmanS., Democratic Co-Learning, 16th IEEE Int Conf Tools with Artif Intell, 2004, pp. 594–602.
31.
LingC.X., DuJ. and ZhouZ.-H., When does Co-training work in real data?Adv Knowl Discov Data Mining, Proc5476 (2009), 596–603.
32.
WangS., WuL., JiaoL. and LiuH., Improve the performance of co-training by committee with refinement of class probability estimations, Neurocomputing136 (2014), 30–40.
33.
KunchevaL.I., Using diversity measures for generating error-correcting output codes in classifier ensembles, Pattern Recognit Lett26(1) (2005), 83–90.
34.
JiangZ., ZhangS. and ZengJ., A hybrid generative/discriminative method for semi-supervised classification, Kn-Based Syst37 (2013), 137–145.
35.
GuoT. and LiG., Improved tri-training with unlabeled data, 2 (2012), 139–147.
36.
DengC. and GuoM.Z., A new co-training-style random forest for computer aided diagnosis, J Intell Inf Syst36(3) (2009), 253–281.
37.
WangJ., LuoS. and ZengX., A random subspace method for co-training, in pp, 2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence) (2008), 195–200.
38.
YaslanY. and CataltepeZ., Co-training with relevant random subspaces, Neurocomputing73(10–12) (2010), 1652–1661.
39.
HallM., FrankE., HolmesG., PfahringerB., ReutemannP. and WittenI.H., The WEKA data mining software, ACM SIGKDD Explor Newsl11(1) (2009), 10.
40.
Alcalá-FdezJ., FernándezA., LuengoJ., DerracJ., GarcíaS., SánchezL. and HerreraF., KEEL data-mining software tool: Data set repository, integration of algorithms and experimental analysis framework, J Mult Log Soft Comput17(2–3) (2011), 255–287.
41.
DengC. and GuoM.Z., Tri-training and data editing based semi-supervised clustering algorithm, Micai 2006 Adv Artif Intell Proc4293 (2006), 641–651.
42.
WangY., XuX., ZhaoH. and HuaZ., Semi-supervised learning based on nearest neighbor rule and cut edges, Knowledge-Based Syst23(6) (2010), 547–554.
43.
CalvoB. and SantaféG., scmamp: Statistical Comparison of Multiple Algorithms in Multiple Problems, 2015, pp. 1–10.
44.
KasperK., GentryJ., LongL., GentlemanR., FalconS., HahneF. and SarkarD., Rgraphviz: Provides plotting capabilities for R graph objects R package ver. 2.15, 2016.
45.
AndrecutM., Parallel GPU Implementation of Iterative PCA Algorithms, 2008.
