tqsar: Random split for glmnet
In angy89/hyQSAR: Hybrid QSAR model

Computes random split validation for glmnet, produces a plot, and returns a value for lambda

tqsar(X_train, X_test, y_train, y_test, type.transformation = "none",
  alpha_values = c(0.1, 0.25, 0.5, 0.75, 1, 1.25, 1.5, 1.75, 2), cost = 3,
  bb = 2, type.solution = "lci", measure = "MSE", intercept = TRUE,
  th = 1.96, cor_th = 0.95, zero_th = 0.8, n.splits = 10,
  inner.train.prop = 0.9, nLambda = 100, n.splits.val = 25,
  n.splits.scr = 25, nCores = 10, md_list = c(), gene_list = c())

`type.transformation`	is a string indicating the type of transformation to be performed at the data before fitting the model. It can be one of the following: "none","abs","abs.alpha.power","mul","log.abs","log.abs.alpha.power"; none is the default parameter.
`alpha_values`	is a numeric vector of alpha value to be used in the abs.alpha.power transformation. Default value is c(0.1,0.25,0.5,0.75,1,1.25,1.5,1.75,2)
`cost`	is a numeric constant used when type.transformation = mul.
`bb`	is the base of the logarithm used when type.transformation = log.abs or log.abs.alpha.power
`type.solution`	is a string indicating the type of solution to compute. Possible values are: min, uci and lci; if standard min or max of the average CV function. if uci, take the the most parsimonous solution within the tot percentage of confidence bands around the standard solution. if lci a less partimosious solution within the tot percentage of confidence bands around the standard solution is selected
`measure`	is the measure used to perform the choice of the optimal lambda value. Possible values are MSE and R2. Default value is MSE
`intercept`	is a boolean valus indicating if we want to fit or not the intercept. Default valuw is TRUE
`th`	is the size of the confidence interval. Default value is 1.96
`cor_th`	is the maximum accepted correlation between couple of features. Default value is 0.95
`zero_th`	is the maximum percentage of zeros accepted in a feature. Default value is 0.8
`n.splits`	is the number of random split to be performed. Default value is 25
`inner.train.prop`	is the percentage of samples from the train test to be used as training set in the random-split method. Default value is 0.9 (90)
`nLambda`	number of lambda to be tested in the LASSO model
`n.splits.val`	is the number of random split to compute validation metrics. Default value is 25
`n.splits.scr`	is the number of random split to perform the y-scrambling test. Default value is 25
`nCores`	is the number of cores to be used
`X`	is the dataset matrix with samples on rows and features on columns.
`y`	is the numeric vector of response variables.

an object of class tqsar containing a list of final models, a list of williams plot object a dataframe with all the metrices computed for the different models, and a list of lambda values used to train the LASSO model.

`finalModels`	a list with one or more models coming from the RCVLasso function
`williams_plots`	a list of one or more williams plot objects
`Metrics`	a dataframe with all the internal and external metrics estimated for every model
`lambda`	a list of one or more vector of lambda used to tune the LASSO models