CAJ | 학술논문

半监督多视图学习是机器学习领域一种极具潜力的大数据处理和分析方法,该方法能有效处理异构和半监督数据,并能方便地在线化和并行化,适合处理海量数据.该方法在大数据时代的应用前景值得研究人员和业界关注.指出未来需要通过引入其他领域新的研究技术和成果,不断丰富和完善半监督多视图学习的理论体系和算法设计,并在实验和实践中不断检验和探索.
반감독다시도학습시궤기학습영역일충겁구잠력적대수거처리화분석방법,해방법능유효처리이구화반감독수거,병능방편지재선화화병행화,괄합처리해량수거.해방법재대수거시대적응용전경치득연구인원화업계관주.지출미래수요통과인입기타영역신적연구기술화성과,불단봉부화완선반감독다시도학습적이론체계화산법설계,병재실험화실천중불단검험화탐색.
This paper introduces a promising machine-learning paradigm cal ed semi-supervised multi-view learning. With this paradigm, information is extracted from heterogeneous and semi-supervised data sets. Lately, multi-view learning has been scaled up online and through paral elization to deal with emerging big data chal enges. Due to its successful application in many research domains and the fact that it has been explored and used by leading companies, multi-view learning may have a future in the big-data era as a major data analytic technique. New research techniques should be introduced into this area to improve the theoretical system and algorithm design of semi-supervised multi-view learning.