Sentiments or emotions are a very important part of human interaction. We can convey our emotions through various means as humans. Some of these means could be through text, speech, gestures, facial expressions etc. Facial Sentiment Analysis, uses the concept of both facial recognition and sentiment analysis to classify sentiments based on facial expressions. It consists of associating an emotion to a facial expression. This can be done using deep learning and convolutional neural networks(CNN). In this paper we will use CNN architecture to classify facial emotions into three categories: Happy, Sad, and Neutral; based on facial features. Comparing our Neural Network with two other pre-trained networks, we find an increase in accuracy by 8% with significantly less parameters than DenseNet. 

The territory of Guam with a population of 156,836 experienced 355 deaths due to COVID-19. Out of these 355, 32% of these deaths were reported Dead on Arrival (DOA). This is comparable to Guam's normal 8% DOA rate and found unusual due to its 88% COVID vaccination uptake level for those who are eligible (>=5). A study was conducted to understand key factors and risk that may influence these DOA statuses. The study interviews families and medical notes to inform the community about potential high risk individuals. Outcomes find that the leading co-morbidity based on odds ratios were diabetes and chronic heart disease which correlate with elevated rates among Guam's population

An outbreak of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), epi-centred in Hubei Province of the People’s Republic of China, quickly spread worldwide and caused COVID-19 pandemic. It infected hundreds of millions of people and caused millions of deaths. In this paper, we develop a compartmental ODE model of COVID-19 transmission. We consider a possibility of breakthrough infections after the vaccination and account for both symptomatic and asymptomatic infections and transmissions. We also incorporate game theory to study the optimal vaccination decisions from the individuals’ perspective. We show that vaccination alone is unlikely to eliminate COVID-19. To achieve herd immunity, the individuals would have to receive a dose of a vaccine more frequently than once every 3 months. It is therefore crucial to adhere to various guidelines, such as quarantine, isolate and wear a mask if tested positive for COVID-19.

Using Statistical Methods, we aim to reevaluate the situation of Guam's Storm Return Periods. The study utilized collaboration between UOG Water and Environment Research Institute (WERI) along with the National Weather Service, to understand dynamics of tropical cyclones in the region and how it affects the island due to rainfall and flooding. We analyze daily rainfall data to better predict the future of rainfall to help evaluate our storm drain systems. This research utilizes the Generalized Extreme Value Family and Typhoon Track to analyze rainfall. Our study produces a schedule and number of days to optomize construction project flow also give probabilities of possible storms.

Game theory is now routinely applied to quantitatively model the decision making of individuals presented with various voluntary actions that can prevent a given disease. Most models consider only a single preventive strategy and the case where multiple preventative actions are available is severely understudied. In our paper, we consider a very simple SI compartmental model of rabies in the domestic dog population. We study two choices of the dog owners: to vaccinate their dogs or to restrict the movements of unvaccinated dogs. We analyze the relatively rich patterns of Nash equilibria (NE). We show that there is always at least one NE at which the owners utilize only one form of prevention. However, there can be up to three different NEs at the same time: two NEs at which the owners use exclusively only the vaccination or movement restriction, and the third NE when the owners use both forms of prevention simultaneously. However, we also show that, unlike the first two types of NEs, the third kind of NE is not convergent stable.