We consider here a large-scale social network with a continuous response observed for each node at equally spaced time points. The responses from different nodes constitute an ultra-high dimensional vector, whose time series dynamic is to be investigated. In addition, the network structure is also taken into consideration, for which we propose a network vector autoregressive (NAR) model. The NAR model assumes each node's response at a given time point as a linear combination of (a) its previous value, (b) the average of its connected neighbors, (c) a set of node-specific covariates and (d) an independent noise. The corresponding coefficients are referred to as the momentum effect, the network effect and the nodal effect, respectively. Conditions for strict stationarity of the NAR models are obtained. In order to estimate the NAR model, an ordinary least squares type estimator is developed, and its asymptotic properties are investigated. We further illustrate the usefulness of the NAR model through a number of interesting potential applications. Simulation studies and an empirical example are presented.

We consider here a large-scale social network with a continuous response observed for each node at equally spaced time points. The responses from different nodes constitute an ultra-high dimensional vector, whose time series dynamic is to be investigated. In addition, the network structure is also taken into consideration, for which we propose a network vector autoregressive (NAR) model. The NAR model assumes each node's response at a given time point as a linear combination of (a) its previous value, (b) the average of its connected neighbors, (c) a set of node-specific covariates and (d) an independent noise. The corresponding coefficients are referred to as the momentum effect, the network effect and the nodal effect, respectively. Conditions for strict stationarity of the NAR models are obtained. In order to estimate the NAR model, an ordinary least squares type estimator is developed, and its asymptotic properties are investigated. We further illustrate the usefulness of the NAR model through a number of interesting potential applications. Simulation studies and an empirical example are presented.

No abstract

This study aimed to examine the relationship between physical performance and mild cognitive impairment (MCI) in Chinese older adults. Methods: The sample comprised 956 relatively healthy and aged ≥65 years old Chinese community-dwelling participants (mean age, 72.56 ± 5.43 years; 56.8% female), which did not include those with dementia, severe cognitive impairment, mental illness etc. The Mini-Mental State Examination (MMSE) and the Instrumental Activities of Daily Living (IADL) scale were used for the initial classification of patients with MCI. Physical performance was measured via hand grip, Timed Up and Go Test (TUGT), and 4-m walking speed. Results: The physical performance (grip strength, TUGT, and 4-m walking speed) correlated with MCI. The grip strength [odds ratio (OR) = 0.96, 95% confidence interval (CI) = 0.93-0.99] and 4-m walking speed (OR = 0.25, 95% CI = 0.10-0.64) correlated negatively with MCI, while TUGT (OR = 1.08, 95% CI = 1.03-1.13) and MCI correlated positively. Conclusion: The physical performance (grip strength, TUGT, and 4-m walking speed) correlated with MCI. Further analysis showed that the grip strength was associated with overall cognition, time orientation, recall, and language, while TUGT and 4-m walking speed were associated with overall cognition and various cognitive domains, except recall.

Raman spectroscopy has been proved to be a promising diagnostic technique for various cancers detection. A major drawback for its clinical translation is the intrinsic weakness of Raman effects. Highly sensitive equipment and optimal measurement conditions are generally applied to overcome this drawback. However, these equipment are usually bulky, expensive and may also be easily influenced by surrounding environment. In this preliminary work, a low-resolution fiber-optic Raman sensing system is applied to evaluate the diagnostic potential of Raman spectroscopy to identify different bladder pathologies ex vivo. A total number of 262 spectra taken from 32 bladder specimens are included in this study. These spectra are categorized into 3 groups by histopathological analysis, namely normal bladder tissues, low-grade bladder tumors and high-grade bladder tumors. Principal component analysis fed artificial neural network are used to train a classification model for the spectral data with 10-fold cross-validation and an overall prediction accuracy of 93.1% is obtained. The sensitivities and specificities for normal bladder tissues, low-grade bladder tumors and high-grade bladder tumors are 88.5% and 95.1%, 90.3% and 98%, and 97.5% and 96.4%, respectively. These results demonstrate the potential of using a low-resolution fiber-optic Raman system for in vivo bladder cancer diagnosis.