keras_regression: Run keras tensorflow regression
In RJ333/phyloseq2ML: Connects Phyloseq To Ranger And Keras

This functions calls keras tensorflow using the parameter values in each row of the provided master_grid, using the data of the list elements. Please have a look at the keras [fit doc](https://keras.rstudio.com/reference/fit.html) for explanation on the keras related variables, the arguments are beginning with "keras" in the description. Except for 'the list', 'master_grid' and '.row' all arguments need to be column names of 'master_grid'

keras_regression(
  Target,
  ML_object,
  Cycle,
  Epochs,
  Batch_size,
  k_fold,
  current_k_fold,
  Early_callback,
  Delay,
  step,
  the_list,
  master_grid,
  .row,
  ...
)

`Target`	factor, the response variable
`ML_object`	factor or char, the name of the corresponding 'the_list' item
`Cycle`	integer, the current repetition
`Epochs`	keras, integer, how many times should the whole data set be passed through the network?
`Batch_size`	keras, integer, how many samples before updating the weights?
`k_fold`	integer, the total number of k_folds for cross validation
`current_k_fold`	integer, the current k_fold in range 1 : k_fold
`Early_callback`	keras, string, a callback metric
`Delay`	keras, integer, wait for how many epochs before callback happens?
`step`	character declaring 'training' or 'prediction'
`the_list`	The input tables list
`master_grid`	the data frame containing all parameter combinations
`.row`	current row of master_grid
`...`	additional features passed by pmap call