(a) Can unlabeled HSI pixel data supplement the limited labeled data? (b) How do resampling techniques work in the SSL process? (c) Can the proposed method provide accurate information when dealing with small-scale and imbalanced spectral data? (d) Can this method be an efficient, rapid, ...
In order to solve the problem of imbalanced data distribution on large-scale network intrusion detection systems, this paper proposes a CSK-CNN model that combines two-layer CNN and imbalanced dataset processing algorithm Cluster-SMOTE + K-means. This paper verifies the anomaly detection rate of th...
Pre-training via unsupervised learning can augment the performance of the model, especially when labeled data are in paucity [59]. In determining the final architecture, we used the Python and Wolfram Mathematica programming platforms. The design process involved iterative ablation experimentation, which...
Guo, H.; Viktor, H. Learning from imbalanced data sets with boosting and data generation: The databoost-im approach.SIGKDD Explor.2004,6, 30–39. [Google Scholar] [CrossRef] Sirisathitkul, Y.; Thanathamathee, P.; Aekwarangkoon, S. Predictive Apriori Algorithm in Youth Suicide Prevention...