Predictive_Power: Predictive_Power

In mbeccuti/PGS: PGS-Peculiar Genes Selection: a feature selection procedure for binary classification tasks.

Description

Predictive_Power is used to fit logistic regression model using DEG gene expression to predict the class labels of the samples. It creates two outputs: a distribution of predictive power associated to each DEG and a Binary Matrix of predicted class labels.

Usage

Predictive_Power(X, y, intercept = TRUE, pp_thr, q_thr)

Arguments

X	Matrix containing predictors values.
y	True classification vector.
intercept	Logical. To include the intercept in the regression model. The default setting is TRUE.
pp_thr	The threshold to switch from the fitted probabilities to predicted class labels. The default value is 0.5
q_thr	The quantile of the Predictive Powers distribution to be chosen as 'Good Predictors'. The defaul setting is 0.95

Value

Predictive_Power returns two files: Predictors_pp and Good_Predictors.

Predictors_pp contains all the predictors names with the associated predictive power. Good_Predictors is the binary matrix containing the class labels of the samples predicted by the DEGs belonging to the selected quantile (qq_thr) of the Predictors_pp distribution.

Author(s)

Federica Martina

References

Liao J, Chin KV. Logistic regression for disease classification using microarray data: model selection in a large p and small n case. Bioinformatics. 2007;23(15):1945–1951.

Van't Veer, Laura J., et al. "Gene expression profiling predicts clinical outcome of breast cancer." nature 415.6871 (2002): 530-536.

Examples

## The function is currently defined as

Predictive_Power('Matrix_Genes', 'classification_vector', intercept = T, pp_thr = 0.6, q_thr = .95)

mbeccuti/PGS documentation built on May 23, 2019, 9:34 a.m.