A neural network for the molecular discrimination of pulmonary arterial hypertension (PAH) and pulmonary veno-occlusive disease (PVOD)

Published: 16 June 2020| Version 3 | DOI: 10.17632/h9nw4cxcs8.3
Contributor:
Helge Stark

Description

This feed-forward neural network for discriminating PAH from PVOD is based on the R packages 'caret' and 'nnet'. It was trained on transcriptomics data acquired with the NanoString nCounter technology. Please see the referenced article for further information! The .rds file contains the serialized model. In order to read the model into an R environment the following steps have to be performed: - Start an R terminal - Execute the 'readRDS' function with the (relative) path to the RDS file as only option and store the returned object in a variable - The returned object is of the class caret::train and can directly be used for the classification of samples (given that the data has been identically prepared/normalized) Please see the manuals of the R packages 'caret' and 'nnet' for help on how to use the loaded objects. Examplary R commands: model.caret <- readRDS(file="model.rds") # object of class 'train' (R package 'caret') containing the final model and all training parameters model.final <- model.caret$finalModel # the final model as object of class 'nnet' (R package 'nnet') print(model.caret$trainingData) # display the data used for training the model

Files

Institutions

Medizinische Hochschule Hannover

Categories

Bioinformatics, Machine Learning, Pulmonary Hypertension, Health, Modelling

License