sine_data_train: Sine Classifier Training Data

In goodekat/limeaid: Diagnose LIME Explanations

Description

A dataset that contains the training portion of the sine classifier data.

Usage

sine_data_train

Format

A data.frame.

Details

The sine classifier data contains 600 observations with three features and one response variable. The features, x_1, x_2, and x_3, were randomly sampled from Unif(-10, 10), Unif(-10, 10), and N(0,1) distributions, respectively. A binary response variable y was created using a rotated sine curve. In particular, let x'_1=x_1\cos(θ)-x_2\sin(θ) and x'_2=x_1\sin(θ)+x_2\cos(θ) where θ=-0.9. Then y is defined as

y=\begin{cases} 1 & \mbox{ if } x'_2 > \sin≤ft(x'_1\right) \\ 0 & \mbox{ if } x'_2 ≤ \sin≤ft(x'_1\right). \end{cases}

500 observations were randomly selected from the generated data to be included in the training portion of the data.

goodekat/limeaid documentation built on March 26, 2021, 10:45 p.m.