This strongly depends on what you mean by 'Predictive Analytics'.
You can easily make applications with QV that helps you to present/analyze the history data and your forecasts.
In some cases the forecasting can be implemented. It's possible but often is not so easy as QV works with whole tables not with cells. In some cases you could even give the users ability modify you forecasts changing the parameters.
Anyway there is a plenty of useful tools that can help you with predictive analytics
But in the end you usually have to interactively present and share your complex results containing a lot of dimensions, facts and relations and that is where QV can definetly help you.
(Some tools can be integrated with QV, for example: http://community.qlik.com/docs/DOC-2975).