Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Forecasting: Principles and Practice (otexts.org)
160 points by yarapavan on Sept 11, 2018 | hide | past | favorite | 11 comments


The book focuses on classical (statistical) methods of forecasting. In this sense, it provides the fundamental notions needed to deal with practical problems. Real world problems are much more complicated and first of all because of the natural of source data which is not limited by univariate numeric data. In most practical cases the success depends on the ability to extract (manually or automatically) good features from heterogeneous data sources. There exist the following frameworks for that purpose:

o https://github.com/asavinov/lambdo - Combines feature engineering and data mining with strong focus on time series analysis

o https://github.com/blue-yonder/tsfresh - Automatically extract informative features (also from time series)


The feature stuff works for time series classification. But still doesn't help with forecasting more than one step ahead.


FYI the authors of this book are authors of the "forecast" R package (used in the book, but also popular in general)

https://github.com/robjhyndman/forecast



What would you all recommend as a good book in forecasting and time series analysis using Python?


Not a book but a nice article: https://tomaugspurger.github.io/modern-7-timeseries

He points to books (including this one) as well as python libraries in the "Resources" section at the end.


This is one of those times when R simply outshines Python. Statisticians mostly use R, so they teach with R and write books that use R and create R packages when they develop new methods. It's why it's worth knowing both.


There isn't really one.


One of the authors, Prof. Hyndman, was also a moderator on stats.stackexchange.com .

See

https://stats.stackexchange.com/users/159/rob-hyndman


Do people rate the Facebook prophet library for time series forecasting?


`prophet` is a really good library for time series forecasting. It's especially useful when speed matters where you create an Rshiny forecasting tool. It's not always the best result though. Sometimes a simple exponential smoothing could give a better result too.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: