relief: RELIEF Feature Selection
In dprep: Data Pre-Processing and Visualization Functions for Classification

Description Usage Arguments Details Value Author(s) References Examples

This function implements the RELIEF feature selection algorithm.

1	relief(data, nosample, threshold,repet=1)

`data`	The dataset for which feature selection will be carried out
`nosample`	The number of instances drawn from the original dataset
`threshold`	The cutoff point to select the features
`repet`	The number of repetitions. It is recommended to use at most 10 repetitions

The general idea of this method is to choose the features that can be most distinguished between classes. These are known as the relevant features. At each step of an iterative process, an instance x is chosen at random from the dataset and the weight for each feature is updated according to the distance of x to its Nearmiss and NearHit. The dataset must have complete cases therefore imputation must be performed in advance.

`relevant`	A table that gives the ratio between the frequency with which the feature was selected as relevant and the total number of trials performed in one column, and the average weight of the feature in another.
`a plot`	A plot of the weights of the features

Edgar Acuna

KIRA, K. and RENDEL, L. (1992). The Feature Selection Problem : Traditional Methods and a new algorithm. Proc. Tenth National Conference on Artificial Intelligence, MIT Press, 129-134.

KONONENKO, I., SIMEC, E., and ROBNIK-SIKONJA, M. (1997). Overcoming the myopia of induction learning algorithms with RELIEFF. Applied Intelligence Vol7, 1, 39-55.

1
2
3

##---- Feature Selection ---
data(iris)
relief(iris,150,0.01,repet=1)

Warning messages:
1: In rgl.init(initValue, onlyNULL) : RGL: unable to open X11 display
2: 'rgl_init' failed, running with rgl.useNULL = TRUE 
3: .onUnload failed in unloadNamespace() for 'rgl', details:
  call: fun(...)
  error: object 'rgl_quit' not found 
Features appearing in at least half of repetitions ordered by their average relevance weight: 
     feature frequency    weight
[1,]       4         1 0.3497222
[2,]       3         1 0.3270056
[3,]       2         1 0.1613889
[4,]       1         1 0.1273148
selected features 
[1] 4 3 2 1