StevenM

This study investigates the role of online discussion forums in facilitating knowledge transfer within an online classroom. Analyzing 21 course sections of a graduate organizational behavior course using NVivo10, the research found evidence of student-to-student knowledge transfer but limited evidence of faculty-to-student transfer. The paper proposes a model to evaluate discussion question prompts as knowledge transfer agents, focusing on factors such as trust, creativity, real-world application, and knowledge stewardship. This model aims to improve the design and delivery of online discussion assignments for instructional designers, subject matter experts, and faculty.

This paper compares two techniques for analyzing web forums with radical content: SentiWordNet and SentiStrength. SentiWordNet is a lexical resource that assigns positivity and negativity scores to words based on WordNet, while SentiStrength uses human-designed lexical and emotional terms along with rules for amplification, diminution, and negation. The study presents and discusses the results of applying these methods to selected forums.

This study investigates the impact of business simulation games on students' learning outcomes by analyzing their interactions in online discussion forums. Through qualitative research of forums from five courses at undergraduate and graduate levels, involving 41 student teams and 3,681 messages, the study finds that certain managerial skills positively influence learning outcomes. Key contributions identified include teamwork, decision-making, information processing, reaching agreements, and handling uncertainty. The findings offer instructional and pedagogical insights into optimizing digital technology methods to enhance student motivation and learning, with recommendations for improving game design and monitoring.

Democratization is a dynamic and contentious process, characterized by the ongoing struggle between social movements and political elites. This struggle aims to increase elite accountability to the broader public, shaping and continually transforming democratic institutions. The process also influences individuals and organizations engaged in social change, leading to the development and evolution of social movement cultures, norms, and practices over time.

This paper presents a framework for extracting contexts and answers from online forum discussions, aiming to create a coherent summary and a valuable QA knowledge base. The proposed approach utilizes Conditional Random Fields (CRFs) to detect relevant information from forum threads. Enhancements to the basic framework include Skip-chain CRFs and 2D CRFs, which are designed to better handle the specific features of forum data. Experimental results demonstrate that these techniques offer promising improvements in performance for context and answer extraction.

This study addresses the problem of information overload and disorganization in Massive Open Online Course (MOOC) discussion forums by developing a model to classify and identify threads based on their relevance to course content. The study defines content-related posts as those that provide or seek help on course material or share/comment on relevant resources. A linguistic model was created using manually-coded posts from a statistics MOOC and tested on posts from the same and another statistics course. The results indicated that content-related posts had unique linguistic features, and the model demonstrated strong cross-course reliability with high recall and precision. The number of views and votes did not contribute to effective classification.

This study explores the automatic detection of suicidal ideation in social media posts, specifically on Reddit. It focuses on using deep learning and machine learning approaches to recognize and classify suicidal content. The research employs a combined LSTM-CNN model and compares its performance with other classification models. Results indicate that the LSTM-CNN model, utilizing word embedding techniques, achieves superior classification results for detecting suicide ideation. The findings demonstrate the effectiveness of deep learning architectures in assessing suicide risk in text classification tasks.

This paper investigates social work practitioners' perspectives on discussing the emotional aspects of their work in both formal and informal forums. The study, conducted through questionnaires and interviews in a Scottish local authority, highlights challenges related to supervision, including issues of safety, permissions, managerial approaches, and individual relationships. Practitioners valued informal peer support for its unrecorded, on-demand, and supportive nature. The findings suggest improvements in supervision practices and emphasize the importance of acknowledging emotions in social work.

This study explores the impact of designing online forums according to deliberative principles on democratic outcomes, comparing them with general online citizen discussions. Conducted in 2013 with 70 participants, the experiment used a post-test only, 2×2 factorial design with a control group. The findings indicate that forums designed for deliberation generally foster positive democratic outcomes—such as coherence of opinions, increased efficacy, trust, and civic participation—though not uniformly across all measures. The study also addresses methodological concerns that may affect the validity of the results and suggests implications for future research.

This paper investigates the use and impact of an online discussion forum in a large-scale beginners' Chinese course at the UK Open University. By analyzing both quantitative data from questionnaires and qualitative data from interviews alongside forum interactions, the study explores the nature, patterns, and functions of the forum among distance language learners. Findings reveal that students had a highly positive experience, with the forum serving as a crucial virtual space for support, resource sharing, and communal learning. It fostered a sense of belonging and course cohesion, enhancing the overall enjoyment and effectiveness of the Chinese language learning process.

Online community forums, increasingly popular over the past decade, face challenges in maintaining quality due to the sheer volume of user questions. Manual moderation is impractical, and traditional methods relying on handcrafted features or community feedback are often inadequate. This study introduces a deep learning approach to assess question quality at creation time. Evaluated on the StackOverflow dataset, which includes 60,000 questions categorized by quality, the proposed model achieves a high F1 score of 0.92, demonstrating its effectiveness in real-time quality monitoring.

Web-based technologies, like Moodle, are increasingly used in second language (L2) classrooms. While empirical studies have shown mixed results regarding their effectiveness, qualitative research indicates that many teachers view these technologies positively. Assessing students' perceptions is crucial because their beliefs about the usefulness of technology can significantly impact their motivation and engagement with it.

This study develops a technique to identify radical opinions in hate group web forums by combining machine learning with semantic approaches. It utilizes four types of text features—syntactic, stylistic, content-specific, and lexicon features—along with three classification techniques: SVM, Naïve Bayes, and Adaboost. The approach is tested on postings from two hate group forums, showing promising preliminary results. Cross-validation demonstrates that the technique is stable and adaptable to different timeframes, making it useful for broader sentiment classification tasks.

Online debate forums are key platforms for expressing and discussing opinions, but often suffer from limited user participation on specific issues, which leads to sparse stance data. This study addresses the challenge of stance prediction by leveraging additional information available in forums, such as user arguments, interactions, and biographical details. An integrated model is proposed that uses a regression-based latent factor approach to jointly analyze these types of side information. The model enhances stance prediction for both users with prior engagement (warm-start) and new users (cold-start). Experimental results show that this method performs well in predicting stances at both micro and macro levels.

As software becomes increasingly integral to daily life, software companies must address a broad range of queries from users, particularly in forums. In open source software communities, users often pose questions that include difficulties, propositions, and opinions. To manage these effectively, constructing a Frequently Asked Questions (FAQ) document is a common strategy. An FAQ helps reduce repetitive questions and saves time for forum members by providing answers to frequently asked questions. This study introduces a pioneering, semi-automatic FAQ finding process designed to help forum managers construct effective FAQs. The approach is evaluated through two case studies, demonstrating its effectiveness in improving FAQ construction.