Maokai Lin – 林懋恺

Journals of a curious learner

© All rights reserved.

Maokai Lin – 林懋恺

Journals of a curious learner

© All rights reserved.

Data Science

Data Science

Hierarchical Bayes Models: Handling Sparse Data After Breakdowns

A common data science problem is that even a huge amount of data could become very sparse after broken down along several dimensions. For example, if you group Uber rides into origin-destination pairs within each city, together with hour of day and day of week, you might see single-digit rides in each group, making it hard to train models independently for each group. The hierarchical Bayes model is a solution for tackling such problem by linking models together.
Read More