Wiki

Clone wiki

otolith / Roadmap

Targeted improvements

  • Pending
    • Adding examples to the documentations
    • Support for more types of classifiers
    • updateproj function with better append utility
    • Adding package vignettes (tutorial)
    • For threcv, adding by-class evaluation + support for agglda; feature to set different threshold values for different classes
    • more flexibility in selection of dimension range in pccv and harcv
    • combined optimization for combined model (multiple views)
    • calculate the shape indices in R
    • integrate multiple views combination for otosearch, with priority to shape indices method
  • Partially achieved
    • Wrapper for processing steps for search and predict purposes (partially solved by release 0.2.0 with img2landmark, which simplify the processing)
    • Sliding semi-landmarks integration for otosearch and otopred function (updated in 0.2.1 release, with thanks to kind helps from the creator of Morpho package Stefan)
  • Achieved
    • Support for combining different views in otopred (updated in 0.2.2 release**)
    • Feature of giving percentage of wrongly predicted specimens in iterated k-fold cross-validation (updated in 0.2.1 release**)
    • Model aggregation feature for otopred- prediction based on aggregated model through either majority vote algorithm or averaging algorithm (updated in 0.2.1 release**)
    • Image-processing in R, do away the hassle to connect to ImageJ (updated in 0.2.0 release, see img2landmark)

Hopefully..

Bugs? New Ideas?

We welcome you to report bug issues or propose new ideas at the Issues page or just drop an email to us.


go back to Home

Updated