Fusing Machine Learning in R
The best layer-specific model is used as meta model.
Cobra Meta Learner
Create COBRA Predictions
Create Difference
Create Loss
createTesting
createTestLayer
createTraining
createTrainLayer
createTrainMetaLayer
Create weights for COBRA Predictions
Abstract class Data
extractData
extractModel
fusemlr
Class HashTable
Lrner Class
Model Class
Best specific Learner prediction.
Predict Using COBRA object
predict.Training
Weighted mean prediction.
PredictData Class
Predicting Class
PredictLayer Class
PredictMetaLayer Class
Testing object Summaries
Training object Summaries
Target Class
TestData Class
Testing Class
TestLayer Class
TestMetaLayer Class
TrainData Class
Training Class
TrainLayer Class
TrainMetaLayer Class
upsetplot
Varsel Class
varSelection
The weighted mean meta-learner
Recent technological advances have enable the simultaneous collection of multi-omics data i.e., different types or modalities of molecular data, presenting challenges for integrative prediction modeling due to the heterogeneous, high-dimensional nature and possible missing modalities of some individuals. We introduce this package for late integrative prediction modeling, enabling modality-specific variable selection and prediction modeling, followed by the aggregation of the modality-specific predictions to train a final meta-model. This package facilitates conducting late integration predictive modeling in a systematic, structured, and reproducible way.