Classifying Survey Papers on Large Language Models Using Machine

With the rapid development of research on large language models (LLMs), the number of related survey papers is continuously increasing, posing challenges for scholars and researchers navigating this field. This study aims to apply machine learning techniques to effectively classify survey papers on LLMs, thereby providing researchers with better literature retrieval and analysis tools. I first constructed a diverse feature matrix by integrating text data and class labels from different datasets. Using preprocessing methods such as TF-IDF vectorization and one-hot encoding, I prepared for subsequent model training. In the experiments, I implemented a random forest classifier to analyze the relationship between features and classification labels. Preliminary results indicate that my classification model achieved an accuracy of 26% without parameter tuning, which improved to 31% after hyperparameter optimization. Despite challenges such as class imbalance and feature correlation, my research provides an effective method for the automatic classification of survey papers. Future work will focus on further improving classification accuracy and exploring other machine learning algorithms to expand the applicability of such tasks.