exactboxes: Exactbox-Algorithm.

In HendrikPfaff/Boxdrawing: Implementation of the boxdrawing algorithms in R

Description

This code requires gurobi and the gurobi-library to solve the MIP.

Usage

exactboxes(positiveTraining, negativeTraining, positiveTesting, negativeTesting,
  cSize, maxK, cExpand, timePerProblem = 4500, varType = "C", env = NULL)

Arguments

positiveTraining	Positive training data.
negativeTraining	Negative training data.
positiveTesting	Positive test data.
negativeTesting	Negative test data.
cSize	Number of different weight for negative data point.
maxK	Cluster size.
cExpand	Parameter to control the number of boxes.
timePerProblem	Time that you are willing to invest for a single MIP problem.
varType

Value

A list containing training TP, training FP, testing FTP and testing FP.

Author(s)

Algorithm and Matlab-code by Cynthia Rudin and Siong Thye Go (see References). Implemented in R by Hendrik Pfaff

HendrikPfaff/Boxdrawing documentation built on May 3, 2019, 3:38 p.m.