Dependency installation suggestion

[simonzhaoms]

If mlhub is run on a non-DSVM or non-DLVM, there may be dependency issue.

The required dependency is now correctly captured:

$ ml demo audit
=====================
Predict Audit Outcome
=====================

Below we see the predictions after applying the pre-built decision tree
model to a random subset of a dataset of previously unseen audit case
outcomes. This provides an insight into the expected future performance
of the model.

mlhub: Required dependencies are not installed for this model: 
  ====> dplyr

To configure the dependencies or download the data required for the model:

  $ ml configure <model>

But how to install the dependency would be more helpful if given the output such as:

$ ml demo audit
=====================
Predict Audit Outcome
=====================

Below we see the predictions after applying the pre-built decision tree
model to a random subset of a dataset of previously unseen audit case
outcomes. This provides an insight into the expected future performance
of the model.

mlhub: Required dependencies are not installed for this model: 
  ====> dplyr

To configure the dependencies or download the data required for the model:

  $ ml configure <model>

Or manually install the dependencies:
  $ Rscript -e 'install.packages("dplyr")'

[gjwgit]

Wording might be:

You can manually install the dependencies:
  $ Rscript -e 'install.packages("dplyr")'