Roadmap for version 0.1

[psolymos]

Roadmap for version 0.1

[psolymos]

Questions

• what to do for 2D samples? Have only XY columns, or add Z=0?

[psolymos]

3D plot:

• use FM (pairwise distances) with MDS,
• make a 3D plot with {rgl}
• color edges between landmarks according to >1 or <1 and CI based significance

[psolymos]

Parametric estimation of SigmaK:

• SigmaK estimation + identifiability checks
• use bootstrap results from object to estimate uncertainty for SigmaK
• add a method to evaluate sensitivity to initial values

[psolymos]

Here is a more detailed roadmap for outstanding features and modifications:

Nonparametric estimation

• remove mixed bootstrap, allow choosing reference (num, denom) for global T test
• find alternate naming (not T or F)

Visualization

• find suitable divergent palette, option to change palette
• implement 3D plot

Data checking

• point cloud around landmarks (as in MorphoJ)
• mean form + 1 individual at a time (indiv ID prnted), or use average?

Parametric estimation

• check landmark names when working with pattern matrix
• make non-estimability error optimal (i.e. allow exploration)
• allow parameter names to be numeric (i.e. numbers as characters)

classification/clustering

• use FM based log(T)-distances for HCA/MDS/k-means to see if there are groups
• predetermined groups, where does it belong? Wishart based likelihood distance ranking

Influential landmark

• LOO type assessment, potentially leaving out multiple landmarks and assess global fit / CI?

GM/GDM revisions

• GM/GDM revisions

Docs

• add Rd pages
• updated tutorial (JSS style)