Authors
Discussion
Loading...
Member takes
Implication
Randomized trial examines additive similarity trees compared to spatial representations in data analysis, suggesting theoretical advantages.
Key Points
- This research aims to explore additive trees as a model for representing similarity data and to compare them with other methods.
- Developed a computer program, ADDTREE, to construct additive trees.
- Applied ADDTREE to several data sets to demonstrate its effectiveness.
- Compared results of additive trees with multidimensional scaling representations.
- Additive trees show empirical advantages over traditional spatial representations of proximity data.
- The comparison highlights the flexibility of additive trees compared to ultrametric trees.
- The findings indicate that tree representations can better capture the structure of dissimilarity data.
Cite This Study
Sattath et al. (1977) studied this question.