Page 344 - Invited Paper Session (IPS) - Volume 2
P. 344
IPS273 Tomoki Tokuda et al.
Figure 4. Classification of D1, D2 and D3. Panel (A): Distribution of subjects.
Panel (B): A classification tree based on Panel (A).
4. Conclusion
We have developed a novel multiple co-clustering method based on
nonparametric Bayesian mixture models. This method can reveal the
underlying multiple view structures in which co-clustering structures with
different types of distributions are imbedded. Applying this method to a
concatenated dataset of different modalities of depression data, we have
identified a relevant view for subtypes of depression. Further analysis of this
view implies a possible classification of treatment-resistant depression and
non-treatment-resistant depression prior to the onset of SSRI treatment. This
result is useful for further investigation of classification of MDD patients from
the perspective of treatment effect.
References
1. Bishop, C.M. (2006). Pattern Recognition and Machine Learning.
Springer.
2. Gu, Q. & Zhou, J. (2009). Co-clustering on manifolds. In: Proceedings of
the 15th ACM SIGKDD International Conference on Knowledge
Discovery and Data Mining. ACM, 359–368.
3. Guan, Y., Dy, J.G., Niu, D. & Ghahramani, Z. (2010). Variational inference
for nonparametric multiple clustering. In: MultiClust Workshop KDD-
2010.
4. Lazzeroni, L. & Owen, A. (2002). Plaid models for gene expression data.
Statistica Sinica, 12, 61–86.
5. Madeira, S.C. & Oliveira, A.L. (2004). Biclustering algorithms for
biological data analysis: A survey. IEEE/ACM Transactions on
Computational Biology and Bioinformatics (TCBB), 1, 24–45.
331 | I S I W S C 2 0 1 9