@techreport{oai:ipsj.ixsq.nii.ac.jp:00048728, author = {北, 研二 and 佐々木, 稔 and Kenji, Kita and Minoru, Sasaki}, issue = {73(1999-NL-133)}, month = {Sep}, note = {代表的な情報検索モデルであるベクトル空間モデルの次元を離散フーリエ変換により削減する方法を提案する。また、提案した方法を概念に基づく検索モデルである潜在的意味インデキシング法(atent Semantic Indexing; LS)へ適用することを試みる。, In this paper, we propose to use the Discrete Fourier Transform (DFT) for dimensionality reduction of the vector space information retrieval model. The point is to apply DFT to document vectors and use the first several Fourier coefficients as document features. We also apply DFT-based dimensionality reduction to Latent Semantic Indexing (LSI). Instead of performing the Singular Value Decomposition (SVD) on the entire term-document matrix, we show that it is sufficient to perform SVD on a DFT-derived reduced space.}, title = {解散フーリエ変換を用いたベクトル空間モデルの次元削減}, year = {1999} }