15th Annual Conference of the International Speech Communication Association

September 14-18, 2014

Subspace Gaussian Mixture Models for Dialogues Classification

Mohamed Bouallegue, Mohamed Morchid, Richard Dufour, Driss Matrouf, Georges Linarès, Renato De Mori

LIA, France

The main objective of this paper is to identify themes from dialogues of telephone conversations in a real-life customer care service. In order to capture significant semantic content in spite of high expression variability, features are extracted in a large number of hidden spaces constructed with a Latent Dirichlet Allocation (LDA) approach. Multiple views of a spoke document can then be represented with several hidden topic models. Nonetheless, the model diversity due to the multi-model approach introduces a new type of variability. An approach is proposed based on features extracted in a common homogenous subspace with the purpose of reducing the multi-span representation variability. A Gaussian Mixture Model subspace model, inspired by previous work on speaker identification, is proposed for theme identification. This representation, novel for theme classification, is compared with the direct application of multiple topic-model representations. Experiments are reported using a corpus collected in the call center of the Paris Transportation Service. Results show the effectiveness of the proposed representation paradigm with a theme identification accuracy of 78.8%, showing a significant improvement with respect to previous results on the same corpus.

Full Paper

Bibliographic reference.  Bouallegue, Mohamed / Morchid, Mohamed / Dufour, Richard / Matrouf, Driss / Linarès, Georges / Mori, Renato De (2014): "Subspace Gaussian mixture models for dialogues classification", In INTERSPEECH-2014, 1880-1884.