Existing research reveals that the misclassification rate for imbalanced data depends heavily on the problematic areas due to the existence of small disjoints, class overlap, borderline, and rare data samples. In this study, by stacking zero-order Takagi-Sugeno-Kang (TSK) fuzzy subclassifiers on the minority class and its problematic areas in the deep ensemble, a novel deep-ensemble-level-based TSK fuzzy classifier (IDE-TSK-FC) for imbalanced data classification tasks is presented to achieve both promising classification performance and high interpretability of zero-order TSK fuzzy classifiers. Simultaneously, according to the stacked generalization principle, the proposed classifier lifts up oversampling from the data level to the deep ensemble level with a guarantee of enhanced generalization capability for class imbalance learning. In the structure of IDE-TSK-FC, the first interpretable zero-order TSK fuzzy subclassifier is built on the original training dataset. After that, several successive zero-order TSK fuzzy subclassifiers are stacked layer by layer on the newly identified problematic areas from the original training dataset plus the corresponding interpretable predictions obtained by the averaging strategy on all previous layers. IDE-TSK-FC simply takes the classical K-nearest neighboring algorithm at each layer to identify its problematic area that consists of the minority samples and its surrounding K majority neighbors. After randomly neglecting certain input features and randomly selecting the five Gaussian membership functions for all the chosen input features and the augmented feature in the premise of each fuzzy rule, each subclassifier can be quickly obtained by using the least learning machine to determine the consequent part of each fuzzy rule. The experimental results on both the public datasets and a real-world healthcare dataset demonstrate IDE-TSK-FC's superiority in class imbalanced learning.

Radiomic model reliability is a central premise for its clinical translation. Presently, it is assessed using test–retest or external data, which, unfortunately, is often scarce in reality. Therefore, we aimed to develop a novel image perturbation-based method (IPBM) for the first of its kind toward building a reliable radiomic model. We first developed a radiomic prognostic model for head-and-neck cancer patients on a training (70%) and evaluated on a testing (30%) cohort using C-index. Subsequently, we applied the IPBM to CT images of both cohorts (Perturbed-Train and Perturbed-Test cohort) to generate 60 additional samples for both cohorts. Model reliability was assessed using intra-class correlation coefficient (ICC) to quantify consistency of the C-index among the 60 samples in the Perturbed-Train and Perturbed-Test cohorts. Besides, we re-trained the radiomic model using reliable RFs exclusively (ICC > 0.75) to validate the IPBM. Results showed moderate model reliability in Perturbed-Train (ICC: 0.565, 95%CI 0.518–0.615) and Perturbed-Test (ICC: 0.596, 95%CI 0.527–0.670) cohorts. An enhanced reliability of the re-trained model was observed in Perturbed-Train (ICC: 0.782, 95%CI 0.759–0.815) and Perturbed-Test (ICC: 0.825, 95%CI 0.782–0.867) cohorts, indicating validity of the IPBM. To conclude, we demonstrated capability of the IPBM toward building reliable radiomic models, providing community with a novel model reliability assessment strategy prior to prospective evaluation.

In this paper, we propose a blockwise combination of interpretable Takagi-Sugeno-Kang (TSK) fuzzy classifiers to simultaneously achieve high accuracy and concise interpretability. As a special hierarchical fuzzy classifier, the proposed classifier is built in a stacked block-by-block way. Each base building block consists of multiple zero-order TSK fuzzy classifiers, which are simultaneously trained in an analytical manner by using negative correlation learning to enhance the generalization ability of the base building block. For utilizing the stacked generalization principle, a random projection of the outputs from the current base building block is presented to the next base building block together with the current training sample in order to enhance the generalization ability of our hierarchical fuzzy classifier. The point of such a special hierarchical structure is that all base building blocks can be trained in the same input-output space with the current training sample and the randomly projected output from the previous building block. In the input layer, the target output for the current training sample is used instead of the randomly projected output from the previous building block. Each TSK fuzzy classifier in base building blocks consists of interpretable TSK fuzzy rules, which are generated by randomly selecting input features and randomly assigning an antecedent fuzzy subset from a fixed fuzzy partition to each of the selected input features. Merits of the proposed classifier are demonstrated through comparative studies on benchmark datasets.

Background: Using high robust radiomic features in modeling is recommended, yet its impact on radiomic model is unclear. This study evaluated the radiomic model’s robustness and generalizability after screening out low-robust features before radiomic modeling. The results were validated with four datasets and two clinically relevant tasks.Materials and methods: A total of 1,419 head-and-neck cancer patients’ computed tomography images, gross tumor volume segmentation, and clinically relevant outcomes (distant metastasis and local-regional recurrence) were collected from four publicly available datasets. The perturbation method was implemented to simulate images, and the radiomic feature robustness was quantified using intra-class correlation of coefficient (ICC). Three radiomic models were built using all features (ICC > 0), good-robust features (ICC > 0.75), and excellent-robust features (ICC > 0.95), respectively. A filter-based feature selection and Ridge classification method were used to construct the radiomic models. Model performance was assessed with both robustness and generalizability. The robustness of the model was evaluated by the ICC, and the generalizability of the model was quantified by the train-test difference of Area Under the Receiver Operating Characteristic Curve (AUC).Results: The average model robustness ICC improved significantly from 0.65 to 0.78 (P< 0.0001) using good-robust features and to 0.91 (P< 0.0001) using excellent-robust features. Model generalizability also showed a substantial increase, as a closer gap between training and testing AUC was observed where the mean train-test AUC difference was reduced from 0.21 to 0.18 (P< 0.001) in good-robust features and to 0.12 (P< 0.0001) in excellent-robust features. Furthermore, good-robust features yielded the best average AUC in the unseen datasets of 0.58 (P< 0.001) over four datasets and clinical outcomes.Conclusions: Including robust only features in radiomic modeling significantly improves model robustness and generalizability in unseen datasets. Yet, the robustness of radiomic model has to be verified despite building with robust radiomic features, and tightly restricted feature robustness may prevent the optimal model performance in the unseen dataset as it may lower the discrimination power of the model.

Brain networks are commonly used to identify cognitive neurobehavioral and brain conscious disorders. Most of the studies on state networks focus on the characterization and expression of resting-state brain networks, but there are few studies on dynamic brain networks. In fact, the analysis of dynamic brain networks can find more valuable information, because it can dynamically depict the dynamic characteristics of brain networks from the time dimension. However, too much consideration of the expression of dynamic networks will naturally hide many of its static characteristics. Therefore, this study proposes a tense-based multi-view fusion fuzzy learning model (TM-FL) to identify dynamic brain networks. TM-FL relearns the features of adjacent networks in high-dimensional space, mining the dynamic features between adjacent brain networks. Best of all, TM-FL can also capture resting information in the brain network at different times in the source space. The fuzzy TSK-based TM-FL also possesses the strong interpretability inherent in fuzzy systems. In this study, a fusion mechanism is designed to effectively integrate decision support under dynamic and static features, and enhance the classification and generalization performance of the proposed fuzzy classification model. For the proposed fusion fuzzy mechanism, Simon entropy is cleverly set as the regular term of the objective function, which not only constrains the decision weight, but also avoids the embarrassment of making mistakes in a leading decision. Experimental results show that TM-FL has good classification, generalization and interpretable capabilities in recognition of multi-channel epileptic electroencephalographic (EEG) signals.

Due to the existence of noise and/or uncertainty in available sleep-wake stage data, recognition of sleep-wake stages with an efficient feature extraction has been facing the following challenges: (1) how to handle these uncertainties; (2) how to have interpretable recognition; (3) how to make recognition of sleep-wake stages to have good testing performance (i.e., generalization capability). To circumvent these three challenges, a deep Takagi-Sugeno-Kang (TSK) fuzzy classifier with random rule heritage called Drrh-TSK-FC is proposed for recognition of sleep-wake stages by integrating enhanced generalization capability of deep structure in fully excavating sleep information hidden in sleep stage data, together with the excellence of fuzzy classifiers in handling uncertainty and providing interpretability in this study. Drrh-TSK-FC assures both enhanced classification performance and strong interpretability by stacking the original input space plus the randomly projected outputs of the previous TSK fuzzy sub-classifier into current sub-classifier in a deep learning way. Besides, it has three distinctive characteristics: (1) the cognitive behavior is mimicked that people often leverage the past experiences (i.e., rules) to solve new yet similar problems. (2) in order to tackle with the changing importance of features caused by the randomness in the generation of the input space for each TSK fuzzy sub-classifier, some randomly generated short fuzzy rules are used. (3) a linear combination of all TSK fuzzy sub-classifiers is taken to facilitate with possibly different importance of the features in each sub-classifier. Moreover, Drrh-TSK-FC can be quickly trained by our previous work—least learning machine (LLM) or its variant in this study. The experimental results on recognition of sleep-wake stage tasks indicate that except for strong interpretability, the proposed Drrh-TSK-FC still keeps at least comparable classification performance to the adopted comparative methods.