SCDV : Sparse Composite Document Vectors using soft clustering over distributional representations

Abstract

We present a feature vector formation technique for documents - Sparse Composite Document Vector (SCDV) - which overcomes several shortcomings of the current distributional paragraph vector representations that are widely used for text representation. In SCDV, word embedding's are clustered to capture multiple semantic contexts in which words occur. They are then chained together to form document topic-vectors that can express complex, multi-topic documents. Through extensive experiments on multi-class and multi-label classification tasks, we outperform the previous state-of-the-art method, NTSG (Liu et al., 2015a). We also show that SCDV embedding's perform well on heterogeneous tasks like Topic Coherence, context-sensitive Learning and Information Retrieval. Moreover, we achieve significant reduction in training and prediction times compared to other representation methods. SCDV achieves best of both worlds - better performance with lower time and space complexity.

10 pages, 5 figures. Update: Added results on Information Retrieval and Topic Coherence with Discussion

References in corpus (5)

• Mixing Dirichlet Topic Models and Word Embeddings to Make lda2vec
• LTSG: Latent Topical Skip-Gram for Mutually Learning Topic Model and Vector Representations
• Representing Documents and Queries as Sets of Word Embedded Vectors for Information Retrieval
• Words are not Equal: Graded Weighting Model for building Composite Document Vectors
• Topic Modeling Using Distributed Word Embeddings

Cited by in corpus (2)

• On-Device Document Classification using multimodal features
• SLAM-Inspired Simultaneous Contextualization and Interpreting for Incremental Conversation Sentences