An explainable sentiment prediction model based on the portraits of users sharing representative opinions in social sensors

"Social sensors" refer to those who provide opinions through electronic communication channels such as social networks. There are two major issues in current models of sentiment analysis in social sensor networks. First, most existing models only analyzed the sentiment within the text but did not analyze the users, which led to the experimental results difficult to explain. Second, few studies extract the specific opinions of users. Only analyzing the emotional tendencies or aspect-level emotions of social users brings difficulties to the analysis of the opinion evolution in public emergencies. To resolve these issues, we propose an explainable sentiment prediction model based on the portraits of users sharing representative opinions in social sensors. Our model extracts the specific opinions of the user groups on the topics and fully considers the impacts of their diverse features on sentiment analysis. We conduct experiments on 51,853 tweets about the "COVID-19" collected from 1 May 2020 to 9 July 2020. We build users' portraits from three aspects: attribute features, interest features, and emotional features. Six machine learning algorithms are used to predict emotional tendency based on users' portraits. We analyze the influence of users' features on the sentiment. The prediction accuracy of our model is 64.88%.

Medical Scribe Impact on Provider Efficiency in Outpatient Radiation Oncology Clinics Before and During the COVID-19 Pandemic

Purpose/Objective(s): Medical documentation has become increasingly challenging for providers, particularly with time constraints and changes to office visit formats during the ongoing COVID-19 pandemic. Medical scribes may help mitigate this burden. Our objective was to determine how scribes affect provider workflow efficiency during the COVID-19 pandemic compared to pre-pandemic controls. Materials/Methods: Providers completed a survey in February 2020 (S1, pre-pandemic) and one year into the COVID-19 pandemic in February 2021 (S2, during-pandemic). Standardized surveys administered during S1 evaluated perceived impact of scribes on clerical work, medical documentation, and efficiency during office visits using the Likert Scale. Surveys administered during S2 also addressed scribe use during telemedicine visits, in addition to office visits. Provider perception of time spent on documentation with or without a scribe was evaluated using a 5-level ordinal scale. Provider response was assessed using descriptive frequency statistics. Fisher's exact test was used to compare categorical variables. Analysis was performed using SAS version 9.4 (SAS Institute Inc, Cary, NC). All tests were two-sided with an alpha level of 0.05. Results: Fifty-eight providers responded to the surveys: 36 (62%) for S1, 22 (38%) for S2. Scribe use decreased perceived clerical work, facilitated chart review, recording of physical exam findings, note documentation and improved efficiency, both before and during the pandemic (P = 0.5, P = 0.7, P = 0.8, P = 0.8, P = 0.9 respectively). Scribe use significantly decreased perceived time to complete documentation pre-pandemic (P = 0.002) and during the pandemic for both in person (P = < 0.0001) and telemedicine visits (P = 0.0004). More providers took over 60 minutes to complete medical documentation without the use of a scribe pre-pandemic (72% versus 30% with a scribe, P = 0.006) and during the pandemic, for both in person (40% versus 0% with a scribe, P = 0.002) and telemedicine visits (35% versus 0% with a scribe, P = 0.002). Even with increased telemedicine visits during the pandemic, 17 (77%) providers strongly agreed that scribe use decreased their daily clerical work and improved efficiency and 18 (82%) strongly agreed scribes were just as helpful during telemedicine visits as during in person visits. Conclusion: Scribe use decreases provider time spent on medical documentation and improves overall efficiency. This improvement in clinical efficiency was similar before and during the COVID-19 pandemic for both in person and telemedicine clinic visits. Integration of scribes into radiation oncology clinics may improve provider satisfaction by reducing burden of documentation and may improve provider well-being.

Data-Driven Modeling to Facilitate Policymaking in Fighting to Contain the COVID-19 Pandemic

This study shows how data-driven modeling can be applied to facilitating policymaking at the geographical hierarchy in terms of the administrative structure of regions and communities when a public health crisis arises. Specifically, rich data and machine learning based models are explored for public health policies, exploring the timing and restrictive levels of intervention measures, such as school/workplace closure and lifting, gathering ban, or travel restrictions, needed for a community, at the region and community level as time goes. This study articulates that rich data and machine learning work well in reducing policy discrepancies. Real world data of COVID-19 cases at the state level in the U.S. are used first in this study to show the consequence of different policy responses in 2020. To demonstrate what different policy responses could result, an agent-based simulation model using a small-scale school setting will be then presented. The simulation model could be further developed, scaled up, and customarily adopted across any geographical hierarchy, facilitating policymaking in public health. (c) 2021 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0) Peer-review under responsibility of the scientific committee of the Complex Adaptive Systems Conference, June 2021.