Fig 1.
The architecture of the Multi-layer CNN model.
Fig 2.
Data collection and analysis effect based on university virtual community.
Table 1.
Mental health evaluation features.
Table 2.
Model parameter settings.
Fig 3.
Multi level CNN model training 3D surface graph.
Table 3.
Training performance of multi-level CNN model under hyperparameter conditions.
Fig 4.
Comparison curve of accuracy between SVM and CNN models.
Table 4.
Comparison of accuracy between SVM and CNN models experimental results.
Fig 5.
Density plot of random forest and CNN model results.
Table 5.
Comparison results of Random Forest (RF) and Convolutional Neural Network (CNN) models.
Table 6.
Comparison of loss degree between GBDT and CNN models during training process.
Fig 6.
Comparison of loss functions between CNN and GBDT.
Table 7.
Comparison of the accuracy of different models.
Fig 7.
Heat map of training loss changes for CNN networks at different levels.
Table 8.
Statistics of training loss changes for CNN networks at different levels.
Fig 8.
t-SNE graph of CNN model before and after data augmentation.
Table 9.
Performance comparison of various models.
Fig 9.
Comparison trend of CNN model accuracy under different loss functions.
Fig 10.
Three dimensional mesh diagram of L2 regularization during CNN model training process.
Fig 11.
Three dimensional heatmap of CNN model training loss under different training epochs.
Fig 12.
Accuracy distribution of CNN models on different college datasets.