摘要
文章研究了基于半监督学习的大数据分类方法,并以MNIST数据集为例,对SGAN模型进行了测试。首先,介绍了半监督学习的基本思想以及SGAN模型的结构,包括标记数据、无标记数据、随机噪声、判别器和生成器等组件。其次,详细描述了在MNIST数据集上测试SGAN模型的方法,包括数据预处理、模型构建、训练过程和性能评估。最后,给出了生成器生成样本的逼真度和多样性指标,以及判别器在测试集上的分类准确率、精确度、召回率和F1分数。经分析验证发现,SGAN模型在MNIST数据集上具有良好性能。
This paper studies the big data classification method based on semi-supervised learning,and tests the SGAN model with MNIST dataset as an example.Firstly,the basic idea of semi-supervised learning and the structure of SGAN model are introduced,including labeled data,unlabeled data,random noise,discriminator and generator.Subsequently,a detailed description was given of the method for testing the SGAN model on the MNIST dataset,including data preprocessing,model construction,training process,and performance evaluation,Finally,the fidelity and diversity indicators of the samples generated by the generator,as well as the classification accuracy,accuracy,recall,and Fl score of the discriminator on the test set,were provided.After analysis and verification,it was found that the SGAN model has good performance on the MNIST dataset.
作者
王少煜
WANG Shaoyu(Yantai Big Data Center,Yantai,Shandong 264000,China)
出处
《计算机应用文摘》
2023年第24期96-98,共3页
Chinese Journal of Computer Application
关键词
大数据分类
半监督学习
半监督生成对抗网络
MNIST数据集
big data classification
semi-supervised learning
semi-supervised generative adversarial network
MNIST dataset
