rPref: Summary of the rPref Package

In patrickroocks/rpref: Database Preferences and Skyline Computation

Summary of the rPref Package

Description

rPref contains routines to select and visualize the maxima for a given strict partial order. This especially includes the computation of the Pareto frontier, also known as (Top-k) Skyline operator, and some generalizations (database preferences).

Preference Composition/Selection

• Preferences are primarily composed from base preferences (see base_pref) and complex preferences (see complex_pref), where especially the Pareto operator for Skylines is such a complex preference.

• Some utility functions for preferences are collected in general_pref.

• Additionally some base preference macros are provided in base_pref_macros.

• The (top(-level)-k) preference selection psel allows to retrieve the maxima of a preference (or Pareto frontier, Skyline), constructed with the functions above, on a given data set.

Visualization and Analysis of Preferences

• The visualization of the preference order in a Better-Than-Graph (Hasse diagram) is possible via plot_btg.

• The adjacency list of the Hasse diagram can be accessed via get_hasse_diag.

• Predecessors/successors in the Hasse diagram are calculated with the pred_succ functions.

• The Pareto frontier can be plotted using the plot_front function.

String Output of Preferences

• The preference query for some preference-supporting DBMS can be given by show.query.

• A preference is partially evaluated and printed with show.pref.

Vignettes

To learn the basics of rPref, start with the vignettes:

• A general introduction and some examples are given in

  vignette("introduction", package = "rPref")

• The visualization of preferences is explained in

  vignette("visualization", package = "rPref")

Further Information

The rPref website is http://www.p-roocks.de/rpref/. To submit bugs, feature requests or other comments, feel free to write a mail to me.

Author(s)

Patrick Roocks, [e-mail]

patrickroocks/rpref documentation built on Feb. 6, 2023, 1:39 p.m.