This post has been de-listed
It is no longer included in search results and normal feeds (front page, hot posts, subreddit posts, etc). It remains visible only via the author's post history.
I am working on a predictive model to predict which applicants are most likely to become students. Example features might include "Visited Campus" and "Paid Deposit". The school year starts only once per year, so when these events happened is important, as is when the prediction is being made.
Is there a way to handle the changing date nature of this type of problem? What should the grain of my data frame be when I build it out? Will I need to build multiple models for different time frames (e.g. "12 months pre-start")?
Subreddit
Post Details
- Posted
- 6 years ago
- Reddit URL
- View post on reddit.com
- External URL
- reddit.com/r/predictivea...