Publication

Abstract : Novelty detection is the process of identifying the presence of new or unknown information in a given text dataset that does not conform to the existing patterns or patterns seen in the past. There have been some studies on the application of spectral graph-based methods for novelty detection in text data. There is still a gap in the literature on the integration of these techniques with visualization methods to enhance the interpretability of the results. Our work aims to fill this gap in the literature by proposing a novel approach that combines spectral graph-based methods with visualization techniques for the task of novelty detection in text data. The proposed approach of combining spectral graph-based methods with visualization techniques can effectively detect novel documents among text data collection while also providing interpretable results. Our contribution is a novel framework that integrates these two methods, which can be applied to various domains that require novelty detection in text data.