###
###  KEEL Banana Data
###

Local copy:
  banana.csv

Downloaded From:
  https://www.openml.org/d/1460
  (CSV file: https://api.openml.org/data/v1/get_csv/1586217)

Original Source:
  KEEL dataset repository
  https://sci2s.ugr.es/keel/dataset.php?cod=182

Variable Information:

  Name -- Data Type -- Measurement -- Description
  V1    -- quantitative -- standardized, unitless -- x coordinate (At1 in the KEEL file)
  V2    -- quantitative -- standardized, unitless -- y coordinate (At2 in the KEEL file)
  Class -- binary (1, 2) -- none -- one of the two banana shapes (coded -1 and 1 in the KEEL file) -- response variable

Summary:
  Number of instances (observations): 5300
  Number of Attributes: 3
  Attribute breakdown: 2 quantitative feature variables, and 1 binary response variable
  Missing Attribute Values: None
  Class distribution: class 1 = 2924 (55.2%), class 2 = 2376 (44.8%)


Abstract:
  An artificial two-dimensional data set whose instances belong to several
  clusters with a banana shape. The class label identifies which of the two
  banana shapes an instance belongs to. The two shapes interleave, so no straight
  line separates the classes; the data is a standard test case for nonlinear
  classifiers such as kernel SVMs.

Data Characteristics:
  Two continuous attributes (the x and y coordinates of each point) and a binary
  class label. The attributes in the OpenML copy are standardized
  (mean 0, standard deviation 1), not in raw units.

Notes:
  Linear methods do poorly on this data. In 5-fold cross-validation (tuned
  hyperparameters), logistic regression reached 56.5% accuracy against a 55.2%
  majority-class baseline, while an RBF-kernel SVM reached 90.4%.


----------
library(tidyverse)
DF <- read_csv("banana.csv")

DF