A Novel Scaled dirichlet-based statistical framework for count data modeling: Unsupervised learning and exponential approximation

Highlights：

• We propose a novel model called the Multinomial Scaled Dirichlet (MSD) for modeling count data.

• We derive a new family of distributions that are approximations to MSD distributions to handle high-dimensional and sparse data that we call (EMSD).

• We develop a minimum message length (MML) criterion for determining of the number of components in EMSD mixture model.

• We evaluate the performance of both approaches through a set of extensive empirical experiments on challenging real-world applications.

• Results revealed that both MSD and EMSD capture the burstiness phenomenon successfully and correctly, and EMSD is many times faster than MSD.

摘要

•We propose a novel model called the Multinomial Scaled Dirichlet (MSD) for modeling count data.•We derive a new family of distributions that are approximations to MSD distributions to handle high-dimensional and sparse data that we call (EMSD).•We develop a minimum message length (MML) criterion for determining of the number of components in EMSD mixture model.•We evaluate the performance of both approaches through a set of extensive empirical experiments on challenging real-world applications.•Results revealed that both MSD and EMSD capture the burstiness phenomenon successfully and correctly, and EMSD is many times faster than MSD.

论文关键词：Count data,Burstiness,DAEM,Multinomial,Scaled dirichlet,Finite mixture models,Exponential family approximation,Model selection,Text collection,Image databases

论文评审过程：Received 3 May 2018, Revised 25 April 2019, Accepted 30 May 2019, Available online 1 June 2019, Version of Record 5 June 2019.