Menu

Post image 1
Post image 2
Post image 3
Post image 4
Post image 5
Post image 6
Post image 7
Post image 8
Post image 9
Post image 10
1 / 10
0

Discrete Time-To-Event Modeling – Predicting When Something Will Happen | Towards Data Science

Towards Data Science·Jarom Hulet·27 days ago
#HyOXS4vx
Reading 0:00
15s threshold

Introduction data science problems predict the what — for example, what will a house sell for? Or what will a customer purchase? Or what is the probability that a patient has a disease? Many real-world decisions, however, depend just as much on when something will happen. How long until a customer churns? When will a loan default? How much time remains before a component fails? Predicting when something will happen is a predictive modeling use case that doesn’t get much attention in introductory materials. Predicting the “when” is often referred to as time-to-event modeling or survival analysis . While event modeling shares techniques and intuitions with more traditional predictive modeling, it also introduces nuances that must be accommodated to create effective predictions. This is the start of a multi-part series that will cover the basics of time-to-event modeling. This first part will discuss basic concepts while future articles will cover time-to-event model development techniques.…

Continue reading — create a free account

Join HashtagPLUS to read full articles, follow hashtags, vote, and join the conversation.

Read More