Modeling Word Relatedness in Latent Dirichlet Allocation
This addresses a limitation in topic modeling for researchers and practitioners, but it is incremental as it builds directly on LDA.
The paper tackled the problem of independent word assignments in standard LDA by proposing WR-LDA, which incorporates word correlation, resulting in improved capabilities like estimating infrequent words and multi-language modeling, with experimental results showing effectiveness compared to standard LDA.
Standard LDA model suffers the problem that the topic assignment of each word is independent and word correlation hence is neglected. To address this problem, in this paper, we propose a model called Word Related Latent Dirichlet Allocation (WR-LDA) by incorporating word correlation into LDA topic models. This leads to new capabilities that standard LDA model does not have such as estimating infrequently occurring words or multi-language topic modeling. Experimental results demonstrate the effectiveness of our model compared with standard LDA.