I am working with healthcare data sets. Mostly to identify patient pathways during their treatment journeys. The solutions be accurate, precise and reasonable for clinicians and managers. Tree like methods are more common sense and can have applications with any type of data.
My objective is to get rule based systems using tree algorithms.
Many thanks for your suggestion/ comments in advance.
Tahseen
Minimum Spanning Trees (MSTs) may be of interest to address overlapped clustering problems:
- Lu et al., " An Effective Algorithm Based on Density Clustering Framework ", 2017 - http://ieeexplore.ieee.org/stamp/stamp.jsp?arnumber=7892858
- Zhong et al., " Minimum spanning tree based split-and-merge: A hierarchical clustering method ", 2011 - http://www.cs.uef.fi/sipu/pub/SplitAndMergeINS2011.pdf
- Cheng et al., " Spatial clustering with Density-Ordered Tree ", 2016 - http://www.homexinlu.com/files/2016%20Lu%20X,%20Physic%20A-Spatial%20clustering.pdf
- Wang et al., " Fast MST-inspired kNN-based Outlier Detection Method ", 2014 - http://booksc.org/book/29176214/24e327
MSTs may be used in a fuzzy context. Refer to:
- Vathy-Fogarassy et al., " Minimal Spanning Tree based Fuzzy Clustering ", 2007 - http://waset.org/publications/1824/minimal-spanning-tree-based-fuzzy-clustering
Gaussian Mixing Models (GMMs) can then be used for deriving a measure of component overlap. Refer to :
- NOWAKOWSKA et al., " TRACTABLE MEASURE OF COMPONENT OVERLAP FOR GAUSSIAN MIXTURE MODELS ", 2014 - https://arxiv.org/pdf/1407.7172.pdf
Finally, an overview of existing methods is available from:
- Ben N’Cir et al., " Overview of Overlapping Partitional Clustering Methods ", 2015
Chapter Overview of Overlapping Partitional Clustering Methods
Evolutionary Algorithms have been proved to be efficient on such problems. A great number of recent papers are published in this domain.
- Badges
- Science topic