Login Sign Up

We get it.
You're asking how the heck this works.
Luckily, it boils down to just a few steps.

1

Modeling

At its core, ClassAI is a prediction powered class browsing and comparison tool. To make that possible, the ClassAI team has developed a cutting-edge prediction model powered by robust deep learning algorithms, which enables the generation of precise slot-based predictions. Any model is only as strong as the data behind it, which in ClassAI's case is 11,000,000 historical snapshots of the registration process and nearly 300,000 historical class sections over a five year period. Accompanying this strong foundation are other datasets, including instructor ratings, student satisfaction scores among other metrics.

2

Extrapolation

To generate precise registration appointment-based predictions, we look to broader related trends to deal with potentially uncovered data points or totally unseen data, such as new courses.

3

Deriving Probabilities

Using our precise point estimates, the momentum of enrollment rates along the registration process, model error, and other factors, we apply logisitic regression to derive a probability that a seat will be open in a given course at a specific registration appointment time.