S
Sutanu Chakraborti
Researcher at Indian Institute of Technology Madras
Publications - 73
Citations - 440
Sutanu Chakraborti is an academic researcher from Indian Institute of Technology Madras. The author has contributed to research in topics: Case-based reasoning & Recommender system. The author has an hindex of 10, co-authored 68 publications receiving 405 citations. Previous affiliations of Sutanu Chakraborti include Indian Institutes of Technology & Tata Research Development and Design Centre.
Papers
More filters
Proceedings ArticleDOI
Document classification by topic labeling
TL;DR: An extension to the Latent Dirichlet Allocation (LDA) based document classification algorithm based on the combination of Expectation-Maximization (EM) algorithm and a naive Bayes classifier is presented.
Proceedings Article
Supervised latent semantic indexing using adaptive sprinkling
Sutanu Chakraborti,Rahman Mukras,Robert Lothian,Nirmalie Wiratunga,Stuart Watt,David J. Harper +5 more
TL;DR: Adaptive Sprinkling is proposed, a more principled extension of LSI that leverages confusion matrices to emphasise the differences between those classes which are hard to separate and can significantly enhance the performance of instance-based techniques to make them competitive with the state-of-the-art SVM classifier.
Book ChapterDOI
Sprinkling: supervised latent semantic indexing
TL;DR: This work proposes an approach that uses class information to influence LSI dimensions whereby class labels of training documents are endoded as new terms, which are appended to the documents.
Book ChapterDOI
Acquiring Word Similarities with Higher Order Association Mining
TL;DR: A weighted linear model is used to combine the contribution of higher orders of co-occurrence into a word similarity model, which outperforms state-of-the-art techniques like SVM and LSI in classification tasks of varying complexity.
Proceedings ArticleDOI
Topic labeled text classification: a weakly supervised approach
TL;DR: An approach that delivers effectiveness comparable to the state-of-the-art supervised techniques in hard-to-classify domains, with very low overheads in terms of manual knowledge engineering is proposed.