stacked autoencoder sae  (Baidu Inc)

Journal: Computational Intelligence and Neuroscience

Article Title: An Improved Stacked Autoencoder for Metabolomic Data Classification

doi: 10.1155/2021/1051172

Figure Lengend Snippet: Fine-tuning structure (SAE, stacked autoencoder).

Article Snippet: In this study, we aimed to introduce an improved framework, named Hessian-free [ ] stacked autoencoder (HF-SAE), combining the Hessian-free algorithm and SAE model with Softmax regression for the classification of metabolomic data of NR-treated RA.

Techniques:

Journal: Computational Intelligence and Neuroscience

Article Title: An Improved Stacked Autoencoder for Metabolomic Data Classification

doi: 10.1155/2021/1051172

Figure Lengend Snippet: Fine-tuning of experimental results on the five-fold data sets. The red and blue lines represent the GD-SAE and HF-SAE results, respectively. In each subgraph of (a) to (e), (i) shows the FMSE, (ii) shows the CR of the training set, and (iii) shows the CR of the test set (GD-SAE, gradient descent stacked autoencoder; HF-SAE, Hessian-free SAE; FMSE, fine-tuning mean square error; CR, classification rate).

Article Snippet: In this study, we aimed to introduce an improved framework, named Hessian-free [ ] stacked autoencoder (HF-SAE), combining the Hessian-free algorithm and SAE model with Softmax regression for the classification of metabolomic data of NR-treated RA.

Techniques: