ExhauFS: exhaustive search-based feature selection for classification and survival regression

Motivation: Feature selection is one of the main techniques used to prevent overfitting in machine learning applications. The most straightforward approach for feature selection is exhaustive search: one can go over all possible feature combinations and pick up the model with the highest accuracy. This method together with its optimizations were actively used in biomedical research, however, publicly available implementation is missing. Results: We present ExhauFS - the user-friendly command-line implementation of the exhaustive search approach for classification and survival regression. Aside from tool description, we included three application examples in the manuscript to comprehensively review the implemented function-ality. First, we executed ExhauFS on a toy cervical cancer dataset to illustrate basic concepts. Then, a multi-cohort microarray and RNA-seq breast cancer datasets were used to construct gene signatures for 5-year recurrence classification. Finally, Cox survival regression models were used to fit isomiR signatures for overall survival prediction for patients with colorectal cancer. Availability: Source codes and documentation of ExhauFS are available on GitHub: https://github.com/s-a-nersisyan/ExhauFS.

COVID-19 Recurrent Varies with Different Combinatorial Medical Treatments Determined by Machine Learning Approaches

Various medical treatments for COVID-19 are attempted. After patients are discharged, SARS-CoV-2 recurring cases are reported and the recurrence could profoundly impact patient healthcare and social economics. To date, no data on the effects of medical treatments on recurrence has been published. We analyzed the treatment data of combinations of ten different drugs for the recurring cases in a single medical center, Shenzhen, China. A total of 417 patients were considered and 414 of them were included in this study (3 deaths) with mild-to-critical COVID-19. Patients were treated by 10 different drug combinations and followed up for recurrence for 28 days quarantine after being discharged from the medical center between February and May, 2020. We applied the Synthetic Minority Oversampling Technique (SMOTE) to overcome the rare recurring events in certain age groups and performed Virtual Twins (VT) analysis facilitated by random forest regression for medical treatment-recurrence classification. Among those drug combinations, Methylprednisolone/Interferon/Lopinavir/Ritonavir/Arbidol led to the lowest recurring rate (0.133) as compared to the average recurring rate (0.203). For the younger group (age 20-27) or the older group (age 60-70), the optimal drug combinations are different, but the above combination is still the second best. For obese patients, the combination of Ribavirin/Interferon/Lopinavir/Ritonavir/Arbidol led to the lowest recurring rate for age group of 20-50, whereas the combination of Interferon/Lopinavir/Ritonavir/Arbidol led to lowest recurring rate for age group of 50-70. The insights into combinatorial therapy we provided here shed lights on the use of a combination of (biological and chemical) anti-virus therapy and/or anti-cytokine storm as a potentially effective therapeutic treatment for COVID-19.

Differentiation of Recurrence from Radiation Necrosis in Gliomas Based on the Radiomics of Combinational Features and Multimodality MRI Images

Purpose. To classify radiation necrosis versus recurrence in glioma patients using a radiomics model based on combinational features and multimodality MRI images. Methods. Fifty-one glioma patients who underwent radiation treatments after surgery were enrolled in this study. Sixteen patients revealed radiation necrosis while 35 patients showed tumor recurrence during the follow-up period. After treatment, all patients underwent T1-weighted, T1-weighted postcontrast, T2-weighted, and fluid-attenuated inversion recovery scans. A total of 41,284 handcrafted and 24,576 deep features were extracted for each patient. The 0.623 + bootstrap method and the area under the curve (denoted as 0.632 + bootstrap AUC) metric were used to select the features. The stepwise forward method was applied to construct 10 logistic regression models based on different combinations of image features. Results. For handcrafted features on multimodality MRI, model 7 with seven features yielded the highest AUC of 0.9624, sensitivity of 0.8497, and specificity of 0.9083 in the validation set. These values were higher than the accuracy of using handcrafted features on single-modality MRI (paired t-test, p<0.05, except sensitivity). For combined handcrafted and AlexNet features on multimodality MRI, model 6 with six features achieved the highest AUC of 0.9982, sensitivity of 0.9941, and specificity of 0.9755 in the validation set. These values were higher than the accuracy of using handcrafted features on multimodality MRI (paired t-test, p<0.05). Conclusions. Handcrafted and deep features extracted from multimodality MRI images reflecting the heterogeneity of gliomas can provide useful information for glioma necrosis/recurrence classification.

Recurrence classification for a family of non-linear storage models

RECURRENCE CLASSIFICATION FOR A FAMILY OF NON-LINEAR STORAGE MODELSNecessary and sufficient conditions for positive recurrence of a discrete-time non-linear storage model with power law dynamics arederived. In addition, necessary and sufficient conditions for finiteness of p-th stationary moments are obtained for this class of models.

Anomalous recurrence properties of many-dimensional zero-drift random walks

Famously, a d-dimensional, spatially homogeneous random walk whose increments are nondegenerate, have finite second moments, and have zero mean is recurrent if d∈{1,2}, but transient if d≥3. Once spatial homogeneity is relaxed, this is no longer true. We study a family of zero-drift spatially nonhomogeneous random walks (Markov processes) whose increment covariance matrix is asymptotically constant along rays from the origin, and which, in any ambient dimension d≥2, can be adjusted so that the walk is either transient or recurrent. Natural examples are provided by random walks whose increments are supported on ellipsoids that are symmetric about the ray from the origin through the walk's current position; these elliptic random walks generalize the classical homogeneous Pearson‒Rayleigh walk (the spherical case). Our proof of the recurrence classification is based on fundamental work of Lamperti.