Spatiotemporal systems are common in the real-world. Forecasting the multi-step future of these spatiotemporal systems based on the past observations, or, Spatiotemporal Sequence Forecasting (STSF), is a significant and challenging problem. Although lots of real-world problems can be viewed as STSF and many research works have proposed machine learning based methods for them, no existing work has summarized and compared these methods from a unified perspective. This survey aims to provide a systematic review of machine learning for STSF. In this survey, we define the STSF problem and classify it into three subcategories: Trajectory Forecasting of Moving Point Cloud (TF-MPC), STSF on Regular Grid (STSF-RG) and STSF on Irregular Grid (STSF-IG). We then introduce the two major challenges of STSF: 1) how to learn a model for multi-step forecasting and 2) how to adequately model the spatial and temporal structures. After that, we review the existing works for solving these challenges, including the general learning strategies for multi-step forecasting, the classical machine learning based methods for STSF, and the deep learning based methods for STSF. We also compare these methods and point out some potential research directions. Machine Learning for Spatiotemporal Sequence Forecasting: A Survey

The first update in a little while brings us release 0.0.4 of RApiDatetime which got onto CRAN this morning via the lovely automated sequence of submission, pretest-recheck and pretest-publish.

Kari Lock Morgan writes:

Deep Learning existing frameworks, disadvantages

In between client work, John and I have been busy working on our book, Practical Data Science with R, 2nd Edition. To demonstrate a toy example for the section I’m working on, I needed scatter plots of the petal and sepal dimensions of the iris data, like so:

Two statistical indices crossed my inbox in the last week, both of which use fast food restaurants to measure a concept indirectly.

R Interface to the Yacas Computer Algebra System (Ryacas)An interface to the yacas computer algebra system.

Watch the second episode of The Dr. Data Show, which answers the question, “How can you trust artificial intelligence?”

I have already written 2 posts about writing functions, and I will try to diversify my content. That said, I won’t refrain from sharing something that has been helpful to me. The function(s) I describe in this post is an artefact left over from before I started using R Markdown. It is a product of its time but may still be of use to people who haven’t switched to R Markdown yet. It is lazy (and quite imperfect) solution to a tedious task.

The principal task of machine learning is to fit a model to some data. Thinking on the level of APIs, a model is an object with two methods: