Semantic Document Clustering Using a Similarity Graph

Abstract

Document clustering addresses the problem of identifying groups of similar documents without human supervision. Unlike most existing solutions that perform document clustering based on keywords matching, we propose an algorithm that considers the meaning of the terms in the documents. For example, a document that contains the words "dog" and "cat" multiple times may be placed in the same category as a document that contains the word "pet" even if the two documents share only noise words in common. Our semantic clustering algorithm is based on a similarity graph that stores the degree of semantic relationship between terms (extracted from WordNet), where a term can be a word or a phrase. We experimentally validate our algorithm on the Reuters-21578 benchmark, which contains 11, 362 newswire stories that are grouped in 82 categories using human judgment. We apply the k-means clustering algorithm to group the documents using a similarity metric that is based on keywords matching and one that uses the similarity graph. We show that the second approach produces higher precision and recall, which means that this approach matches closer the results of the human study.

Survey on Semantic Similarity Based on Document Clustering

“…As [40], the approach for semantic document clustering based on the graph similarity. Graph is mainly used information from WordNet to the degree of semantic similarity between 150,000 of the most common in English terminology.…”

Fine-Tuning an Algorithm for Semantic Document Clustering Using a Similarity Graph

“…A preliminary version of this article was published in the conference proceedings of the Tenth IEEE International Conference on Semantic Computing [40]. Here, the paper is signi¯cantly revised, corrections are made, and more detailed explanations are provided in every section.…”

“…Our approach di®ers because we use supervised learning only for ¯ne-tuning the algorithm. For example, our original algorithm in [40] is unsupervised, it does not use a training set, and it can cluster documents in any number of classes rather than just classify the documents in preexisting categories.…”

“…We ran our algorithm from [38] with di®erent values for a 3 on the training set and the results are shown in Table 4. As the table suggests, the optimal value is when a 3 ¼ 0:3, which is the value that is used in the algorithm from [40].…”

Document Clustering using a Graph Covering with Pseudostable Sets

“…The authors of [3] for example try to cluster documents from MEDLINE by using evolutionary algorithms, whereas [4] use machine learning approaches. Only few authors like [5] use graph-based approaches. Some authors, like [6] cover related problems like clustering in the context of search queries, whereas [7] work on the field of hierarchical clusterings.…”