Clustering short text using Ncut-weighted non-negative matrix factorization

• DOI: 10.1145/2396761.2398615
• Publication Date: 2012-11-15T16:38:00Z
• Authors: Xiaohui Yan; Jiafeng Guo; Shenghua Liu; Xue-qi Cheng; Yanfeng Wang
• Keywords: Non-negative Matrix Factorization; Discriminative model; tf–idf; Document Clustering

Non-negative matrix factorization (NMF) has been successfully applied in document clustering. However, experiments on short texts, such as microblogs, Q&A documents and news titles, suggest unsatisfactory performance of NMF. An major reason is that the traditional term weighting schemes, like binary weight tfidf, cannot well capture terms' discriminative power importance due to sparsity data. To tackle this problem, we proposed a novel scheme for NMF, derived from Normalized Cut (Ncut) problem affinity graph. Different idf, which emphasizes discriminability level, Ncut measures level. Experiments two data sets show our significantly boosts NMF's text