Abstract:
The Research Paper Acceptance Prediction System leverages machine learning (ML)
and natural language processing (NLP) to predict the likelihood of research paper acceptance
at academic conferences. The traditional peer review process is often subjective, time
consuming, and prone to inconsistencies, leading to challenges in ensuring fairness and
transparency. This research aims to address these limitations by developing a predictive
system that evaluates papers based on multiple factors, including content quality, relevance,
novelty, and peer review sentiment.
The system utilizes TF-IDF and BERT-based embeddings for feature extraction,
combined with Logistic Regression and advanced ML models for classification. Additionally,
explainable AI (SHAP/LIME) techniques are integrated to provide justifications for
acceptance or rejection, enhancing interpretability. The project follows a modular design,
ensuring scalability, efficiency, and usability, with a user-friendly interface for paper
submission and feedback generation.
This study contributes to automating and improving the research paper evaluation
process, offering researchers insights into their submissions and constructive feedback for
refinement. The model’s performance is evaluated using accuracy, precision, recall, and F1
score, ensuring robustness and reliability. By bridging the gap between human peer review
and AI-driven analysis, this system has the potential to enhance decision-making in academic
paper selection while maintaining fairness and transparency.
