Clickbait Challenge at SemEval 2023 - Clickbait Spoiling Abstract This study tackles the problem of clickbait by classifying posts into 'phrase,' 'passage,' and 'multipart' categories using the Webis Clickbait Spoiling Corpus 2023, part of SemEval 2023. The original dataset is split into two subsets: 3,200 training posts and 800 validation posts. These entries are formatted in JSON Lines and include fields such as uuid, postText, and targetParagraphs. I applied various data preprocessing techniques and TF-IDF feature extraction, followed by classification using Naive Bayes, Logistic Regression, and Support Vector Machines (SVM) models. The SVM model achieved 75.9% accuracy, while Logistic Regression and Naive Bayes reached 87.9% and 88.2%, respectively. These results highlight the effectiveness of machine learning in distinguishing between different types of clickbait, improving content moderation, and enhancing user experiences by providing more reliable online content. Future research will focus on implementing advanced deep learning models to further refine these methods and expand their applicability across diverse datasets.
ashedrack / course_work_7072cem Goto Github PK
View Code? Open in Web Editor NEWA Machine Learning Approach on Classifying Orthopedic Patients Based on Their Biomechanical Features