fsEnsemble: Ensemble feature selection for MTS

In fsMTS: Feature Selection for Multivariate Time Series

Ensemble feature selection for MTS

Description

fsEnsemble implements methods for ensemble learning of features for multivariate time series

Usage

fsEnsemble(feature.sets, threshold, method = c("ranking", "majority"))

Arguments

feature.sets	a list of matrixes that contains weights for features, estimated by several feature selection algorithms (base learners)
threshold	the required sparsity of the resulting feature set
method	a ensemble learning algorithm. Implemented algorithms: "ranking" - individual feature sets are ranked according to their weights and further the sum of ranks is used for feature selection (threshold share of features is selected). The algorithm uses ranking of feature with a minor priority to earlier lags and even smaller priority to order of MTS components. So, if features of 1st and 2nd lags have identical weights, the feature of the 1st lag will be preferred; if features of the same lag have identical weights, the order of features is used as a priority. "majority" - base feature sets are for feature selection (threshold share of features is selected) and further the resulting feature set is estimated using majority voting (50 or more percent of base learners)

Value

returns a binary feature matrix. Columns correpond to components of the time series; rows correspond to lags.

References

Pes, B., 2019. Ensemble feature selection for high-dimensional data: a stability analysis across multiple domains. Neural Computing and Applications. https://doi.org/10.1007/s00521-019-04082-3

Examples

# Load traffic data
data(traffic.mini)

# Scaling is sometimes useful for feature selection
# Exclude the first column - it contains timestamps
data <- scale(traffic.mini$data[,-1])

mIndep<-fsMTS(data, max.lag=3, method="ownlags")
mCCF<-fsMTS(data, max.lag=3, method="CCF")
mDistance<-fsMTS(data, max.lag=3, method="distance", shortest = traffic.mini$shortest, step = 5)
mGLASSO<-fsMTS(data, max.lag=3,method="GLASSO", rho = 0.05)
mLARS<-fsMTS(data, max.lag=3,method="LARS")
mRF<-fsMTS(data, max.lag=3,method="RF")
mMI<-fsMTS(data, max.lag=3,method="MI")
mlist <- list(Independent = mIndep,
              Distance = mDistance,
              CCF = mCCF,
              GLASSO = mGLASSO,
              LARS = mLARS,
              RF = mRF,
              MI = mMI)


th<-0.30
mlist[["EnsembleRank"]] <- fsEnsemble(mlist, threshold = th, method="ranking")
mlist[["EnsembleMajV"]] <- fsEnsemble(mlist, threshold = th, method="majority")
(msimilarity <- fsSimilarityMatrix(mlist,threshold = th, method="Kuncheva"))

fsMTS documentation built on April 26, 2022, 9:05 a.m.