Minnesota State University, Mankato
Institutional Repository for Minnesota State University, MankatoNot a member yet
33414 research outputs found
Sort by
Comparative Analysis of Data Augmentation on Sentiment Analysis in Three Distinct Languages
Get PDFMachine learning in natural language processing analyzes datasets to make future predictions for various filed in the real world. By training machine algorithms on the datasets of text, the model can learn patterns and structure of the text in many different languages. Then the model enables to perform the text classification, sentiment analysis, and other tasks. A large and balanced dataset is required to develop an accurate machine learning model. However, the collection of a reliable, large, and equally distributed dataset is a challenging and requires significant resources and time. As a solution to this challenge, a data augmentation technique can be used to increase the size of a dataset by generating new data from the original dataset. This study investigates the impact of data augmentation on the performance of a machine learning models using small datasets in three diverse languages: French, German, and Japanese. After the data augmentation inflates the three diverse languages training datasets, three models are trained by each augmented training dataset. The three models’ performance were compared with other three models’ performance that are trained by each three original training datasets. This not only addresses the issue of a lack of large and balanced datasets but also the issue of dataset scarcity in various areas. Towards this, the generalization of each model trained by an augmented dataset is evaluated on each test dataset in different languages. A machine learning’s capability of generalization can contribute situations where cross-lingual capabilities are needed, such as, international market research, multilingual customer support, obtaining cultural insights, etc. The models\u27 performances and generalization are measured through evaluation metrics: accuracy, precision, recall, and f1-scores. Our results show that data augmentation improved the performance of the model’s sentiment analysis with the languages French and Japanese. The results also showed that a model trained with a Japanese dataset showed improved performance in sentiment analysis when tested using German test data and vice versa. Similarly, a model trained with the German dataset showed marked improvement in its performance in sentimental analysis when tested with the French test dataset and vice versa
Religious and Spiritual Struggles among Satanists
Get PDFAs Satanism represents a unique religious demographic, the present study sought to determine the prevalence of religious and spiritual struggles among modern Satanists and resulting anxiety or depressive symptoms. A sample of 693 self-identified Satanists were surveyed about their experiences of religious and spiritual struggles, perceptions of their Satanist identity, and anxiety and depressive symptoms. Results indicated that Satanists predominantly struggle with interpersonal and ultimate meaning struggles. Ultimate meaning and moral struggles were predictive of both anxiety and depressive symptoms. Interpersonal struggles also predicted anxiety symptoms. Individual perceptions of Satanist identity did not have a significant influence on the relationship between R/S struggles and mental health outcomes. Future research directions on Satanism and mental health are discussed
Volume 44, Number 4, December 2024 OLAC Newsletter
Get PDFDigitized December 2024 issue of the OLAC Newsletter
Increased Student Employment is Associated with Inferior Biology Exam and Course Performance
No full textRising tuition rates are detailed as a driver of increased student employment, potentially leading to reduced time and academic performance. We observed this relationship across 3 cohorts of Genetics (BIOL-211), totaling 238 students. Students working 20+ hours scored significantly fewer total points (p=0.0089), exam points (p=0.0255), and were more likely to incur a failed assignment (p=0.0025). These data represent an important metric for identifying and treating underlying factors associated with reduced STEM performance and retention.
NOTE: Video of this presentation begins at 32:00
Understanding Communication Research Methods: A Theoretical and Practical Approach
No full textUsing an engaging how-to approach that draws from scholarship, real life, and popular culture, this textbook, now in its fourth edition, offers students practical reasons why they should care about research methods and offers a practical guide for conducting research.
Explaining quantitative, qualitative, critical, and performance research methods, this new edition helps students better grasp the theoretical and applied uses of method by clearly illustrating practical applications. The book features all the main research traditions in communication, including applications of the methods through effective examples and exercises, and sample student papers that demonstrate research methods in action.
This textbook is perfect for beginning and advanced scholars using critical, cultural, interpretive, qualitative, quantitative, rhetorical, and performance research methods.
Additional resources for students and instructors can be found on the eResource at www.routledge.com/9781032557380, which includes links, videos, outlines, activities, recommended readings, test questions, and more.https://cornerstone.lib.mnsu.edu/university-archives-msu-authors/1481/thumbnail.jp
Meditation in the Woods: Is There a Return Effect?
No full textDecades of empirical evidence suggest that immersive experiences in Nature can promote psychological health. Research similarly demonstrates the ability of contemplative practices such as guided visual imagery to facilitate health and healing. We consequently sought to explore the potential of a synergy between these two lines of evidence. Informed by the combined literature and guided by a previous pilot study, we took twenty-two college professionals to the woods for a five-day immersive experience to leverage the synergistic potential of Nature and meditation to foster improved well-being in a non-clinical eco-therapeutic manner. The results of the study were statistically robust and contextually verified by the responses to a brief set of open-ended questions. We specifically demonstrated that the immersive field experience resulted in (a) improved thinking, (b) a more positive mood, (c) a less negative mood, and (d) a deeper connection with Nature. However, the longer-term results were murkier one month after a return to home and work. The data suggest that the subjects were doing worse across our metrics than they did at the time of pretesting on the first day in camp. We suggest two competing hypotheses, either or both of which might prove explanatory of the follow-up results. Regardless, the results from the field demonstrated a meaningful and statistically significant effect of guided visual imagery immersed in a natural setting on thought, mood, and human growth learned over a mere handful of days
Developing a Snow Detection Algorithm Using Spatial Attention for Pedestrian Safety
Get PDFSNOW-COVERED SIDEWALKS POSE SIGNIFICANT SAFETY HAZARDS, ESPECIALLY FOR VULNERABLE POPULATIONS SUCH AS THE ELDERLY AND VISUALLY IMPAIRED. THE DEVELOPMENT OF EFFECTIVE SNOW DETECTION SYSTEMS IS CRUCIAL FOR ENHANCING PEDESTRIAN SAFETY. THIS RESEARCH AIMS TO ADDRESS THESE CHALLENGES BY DEVELOPING A SNOW DETECTION ALGORITHM SPECIFICALLY DESIGNED FOR SIDEWALKS. THE PROPOSED ALGORITHM USES A CONVOLUTIONAL NEURAL NETWORK (CNN) ARCHITECTURE INCORPORATING A 2-DIMENSIONAL SPATIAL ATTENTION MECHANISM TO FOCUS ON RELEVANT FEATURES IN IMAGES, IMPROVING SNOW DETECTION ACCURACY. DUE TO THE SEASONAL AND GEOGRAPHIC LIMITATIONS OF SNOW DATA COLLECTION, SYNTHETIC DATA GENERATION USING INVERSE DIFFUSION MODELS WAS EMPLOYED TO AUGMENT THE REAL-WORLD DATASET. ALTHOUGH THE INVERSE DIFFUSION-GENERATED IMAGES DID NOT SIGNIFICANTLY IMPROVE DETECTION PERFORMANCE COMPARED TO REAL-WORLD TEST DATA, THEY PROVIDED VALUABLE INSIGHTS INTO THE POTENTIAL FOR SYNTHETIC DATA AUGMENTATION. THE MODEL\u27S PERFORMANCE, EVALUATED THROUGH STANDARD METRICS LIKE RECALL, PRECISION, AND F2 SCORE, CONSISTENTLY SURPASSED ESTABLISHED MODELS SUCH AS FINE-TUNED VGG-19 AND RESNET-50. MOREOVER, THE MODEL DEMONSTRATED FASTER INFERENCE SPEEDS, MAKING IT SUITABLE FOR REAL-TIME APPLICATIONS. THE FINDINGS OF THIS STUDY UNDERSCORE THE IMPORTANCE OF DEVELOPING SPECIALIZED MODELS FOR SNOW DETECTION AND HIGHLIGHT THE ONGOING NEED FOR REFINING SYNTHETIC DATA GENERATION TECHNIQUES TO IMPROVE REAL-WORLD APPLICABILITY
Dance as a Collective Action: Creating Relationships, Communities, and Change
Get PDFDance is a place of belonging—to ourselves, to our homes, lands, peoples, cultures, music and so much more. Dance is in relation—embodying stories of communities and dreaming of shared future. Dance is an action—having the ability to enact change. In this thesis paper, I explore how our dancing bodies can create unique relationalities to each other and build sites of transformation. To situate dance as rooted in community, I propose four sites: the personal body, dance learning space, dance-making, and performance. Through these sites, I examine how dance creates belonging in our personal bodies, it adapts and shifts in response, and it opens pathways toward recognizing and honoring diverse bodies in shared dancing spaces. Dance spaces built on collective care and shared goals generate interconnected webs of relationality amongst diverse embodiments. I reference dance scholars, researchers, artists, and makers to share how dance is a collective action, that creates change and builds connectivity
Volume 47, 2024 Communication and Theater Association of Minnesota Journal
Get PDFComplete digitized volume (volume 47) of Communication and Theater Association of Minnesota Journal
Lexical and Textual Development through Autonomous Engagement with Vocabulary Journal in Multilingual Learners
No full textResearch has demonstrated a positive correlation between the use of diverse and sophisticated vocabulary and the overall quality scores of essays in students\u27 writing (Ferris, 1994; Jarvis et al., 2003). Despite this, the instruction of vocabulary in composition courses has historically been minimal, largely due to prevailing views from the history of second language acquisition theory, which suggest that explicit teaching of linguistic features is unfavorable (Brannon & Knoblauch, 1982; Hartwell, 1985; Krashen, 1982, 1984). This is corroborated by Ferris (2014) and Folse (2004), who note the virtual absence of vocabulary instruction in these contexts. Conversely, the challenge posed by academic-specific vocabulary, especially for L2 learners who cannot intuitively grasp English grammatical rules and notions of \u27correctness\u27, has been acknowledged (Frodesen & Holten, 2003), and the importance of learning vocabulary in context has been emphasized (Laufer & Hulstijn, 2001; Paribakht & Wesche,1999). This study investigated how the creation of an Vocabulary Journal (VJ) based on course readings influences writing within the course. The intervention, involving multilingual university freshmen (N=12), spanned ten weeks during which students extracted vocabulary from scholarly articles used as course readings and created a VJ. The intervention utilized a Vocabulary Journal (VJ) that was developed by incorporating modifications to the models proposed by Ferris and Hedgcock (2014) and Staehr Fenner and Snyder (2017), following a ii review of major theories related to vocabulary acquisition. Research on how English language learners (ELLs) select words from reading texts to create a Vocabulary Journal (VJ) and apply this vocabulary in academic writing has been scarce. This study contributes to the field by examining the characteristics of words recorded in the VJ, tracking the number and methods of their application in subsequent writings over time, and analyzing changes in Lexical Richness Indicators in students\u27 writing as time progresses. Additionally, the study explores correlations with interest in the readings and changes in students\u27 perceptions before and after the intervention. The results are analyzed both quantitatively and qualitatively, offering insights into the dynamics of vocabulary acquisition and usage in academic contexts