prediction of completion time of a work item I have work items given features such as complexity, work item type etc.. I know how long it took to complete a work item given its features in the past. I would like to predict how long it takes to complete a given work item (on average).
Is this a regression problem or could it be a survival analysis problem?
 A: This is a regression problem that can/should use survival analysis as the underlying distribution. I recommend the Weibull or exponential model for ease of creating predictions, if prediction is your main goal. Depending on how your data are structured, the negative binomial distribution may offer a meaningful solution.
A: There are many different techniques you could use to predict how long, on average, a work item will take.  Regression is certainly a method that could be used, but there are also a number of other statistical and machine learning techniques that you could use to predict how long the work task takes.  For example, neural networks can be used to predict the time to complete a task.  As other users have stated, a survival analysis could also be used to predict how long until a task takes until it is completed (in this case "completion" is considered the "failure" in the traditional survival analysis lingo).  That being said, there are certain types of survival analyses that fall under "regression" techniques.  So in this sense, certain survival analyses like Cox-Proporational Hazards models are a specific type of regression.  In this case, the question isn't an either/or.
