Mason Journals (George Mason Univ.)
Not a member yet
3256 research outputs found
Sort by
A Comparative Study of Decision Tree and Random Forest ML Algorithms for Estimating PM2.5 Levels from AOD
No full textGround monitoring systems to measure PM2.5 levels are expensive to deploy in remote regions. A lack of air quality data for such regions impacted by frequent forest fires leads to excessive exposure to smoke, causing increased respiratory diseases in these regions. With a proper machine learning model, the AQI can be estimated from readily available AOD data for such regions. In this paper we compare two machine learning algorithms: Decision Tree and Random Forest (RF), accounting for various locations and conditions. This comparative study aims to select the optimal model based on factors such as temperature, humidity, season, and geography. PurpleAir sensor data is used for training and preprocessing, classification of indoor and outdoor sensor data, averaging of the data where applicable, and grouping based on season. Based on the preliminary analysis, the RF model is up to 50% more accurate when both indoor and outdoor sensor data were used and when the region is impacted by a forest fire. Both algorithms have similar estimates when trained with a homogenous dataset excluding extreme weather events like fire or storm. There is a significant variation in the accuracy due to weather events like rain due to increased humidity. However, the use of a singular dataset to train the model limits its accuracy in predicting various types of wildfires. This paper concludes that while it is possible to estimate PM2.5 close to ground monitoring systems, the training data and model needs to consider all influencing factors to improve accuracy
Keylogging in Virtual Reality: Assessing Data Vulnerabilities via Motion-Position Sensors
No full textThe integration of advanced motion-position sensors in Virtual Reality (VR) systems has significantly enhanced user interaction but has also raised serious privacy concerns. These sensors track the position and orientation of VR controllers, potentially revealing sensitive information about user activities. Existing research has demonstrated that sensor data can lead to privacy attacks, but the methodologies described are often vague and lack critical details. To address these gaps, we developed a comprehensive framework to log VR controller data without requiring user permissions for background operations. We have put significant efforts into understanding and replicating existing methodologies, as well as fixing missing details. Our approach includes identifying typing windows through detailed analysis of trigger presses, allowing for accurate capture and classification of user input. We then estimate 3D cursor positions, which are used to train a K-Nearest Neighbors (KNN) model for sentence reconstruction. Our method achieves a 96.1% accuracy rate in this task. By providing detailed replication methodologies and addressing gaps in existing work, our research benefits the VR security community by ensuring the reproducibility of VR privacy attacks, thereby aiding the development of advanced defense solutions
How reliable is disability-related information on social media? Exploring the prevalence of disability misinformation and assessing fact-checking tools.
No full textSocial media platforms such as YouTube and Facebook are key sources of disability information but are also prone to misinformation. Our study investigates how disability-related misinformation is shared and disseminated through social media. Specifically, we ask the questions: (1) how prevalent is misinformation regarding disabilities? (2) how does the reliability of fact-checking tools look in evaluating the truthfulness of claims? To answer these questions, we collected data from Facebook and YouTube. YouTube data was collected through YouTube APIs using 28 keywords, yielding 13,034 videos with transcripts. Facebook data was collected using EasyScraper from 20 public disability-related pages and groups identified in a large-scale survey. The data, initially consisting of 21,498 posts, was refined with ChatGPT-4o API to identify information both fact-checkable and relevant to disabilities, yielding 1,414 data points. We fact-checked 500 randomly sampled posts by manually cross-referencing reputable sources and using AI tools (Originality.ai and ChatGPT-4o). Manual checks identified 85% of posts as “correct” and 15% as “mixed”, “incorrect”, or “non-objective”. Originality.ai identified 53% as “correct” and 47% as “mixed”, “incorrect”, or “non-objective”. ChatGPT marked 56% as “correct” and 44% as “mixed”, “incorrect”, or “non-objective”. Misinformation had the highest frequency in advertisements, followed by personal anecdotes. AI tools produced different results than ground-truths and overestimated the prevalence of misinformation, highlighting the necessity to choose such tools carefully
World History and the Temple of Time: Reconsidering Emma Willard’s Signature Illustration of Human History
Get PDFThis article suggests that the work of pioneering 19th century educator Emma Willard should be revisited for her contributions to world history, and that her important illustration of human history, which she called the Temple of Time, can serve as an excellent teaching tool in contemporary world history classrooms
Reacting to the Past Games in Introductory World History Courses: How to Easily Revise Your Teaching for Transformative Learning and Have Fun in the Process
Get PDFThis paper discusses the use of Reacting to the Past in the World History Survey, arguing for the value of deep dives into specific periods, getting students to engage in deep learning. In particular, it will make the case for using several shorter games to cause students to evaluate how at different times in world history, different societies engaged in diplomacy, warfare, and negotiation. Weighing options in Warring States China, ancient Athens, and between Dutch settlers and Khoe communities in Southern Africa, this paper emphasizes the lasting effects of Reacting Games, and their ability to cause students to reflect critically on their own assumptions, while re-evaluating the sources and issues they come across in non-game portions of the class. Furthermore, the paper engages in critical approaches to Reacting Games, highlighting the need for more games that center women’s experiences in the emerging global south. 
