Machine Learning Aided Text Clustering

Unsupervised clustering is a popular research domain in machine learning on natural language while supervised methods like intent classification achieve remarkably good results in text data. A latent variable-based model is introduced to improve the clustering results on NLP tasks. We exploit recurrent auto-encoder to obtain latent representations of sentences, then iteratively optimize an objective that learns feature representations and clustering assignments simultaneously. Experiments on several datasets show the need for preserving sequence information and the effectiveness of the proposed method.

I applied the DEC model on sequence to sequence auto-encoder to improve performance on text clustering. Different structures of the recurrent auto-encoder model are also discussed in my experiments. I proposed a model structure which learns a better representation for text data.