search

srlanalytics / bdfm

Bayesian dynamic factor model estimation and predictive statistics including nowcasting and forecasting
MIT License
5 stars 6 forks source link

Seas we() updates #30

Closed srlanalytics closed 5 years ago

srlanalytics commented 5 years ago

This has a conflict with time-series-w-NA but I'll hopefully get the merge sorted out since there are a lot of updates here I want to keep (and the US Nowcasting example).

codecov-io commented 5 years ago

Codecov Report

Merging #30 into master will decrease coverage by 0.24%. The diff coverage is 39.28%.

Impacted file tree graph

@@            Coverage Diff             @@
##           master      #30      +/-   ##
==========================================
- Coverage   41.31%   41.07%   -0.25%     
==========================================
  Files          10       10              
  Lines        1941     1984      +43     
==========================================
+ Hits          802      815      +13     
- Misses       1139     1169      +30
Impacted Files Coverage Δ
R/seas_factors.R 67.85% <100%> (ø) :arrow_up:
R/seas_we.R 54.78% <38.18%> (-14.67%) :arrow_down:

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update 8f729dc...d17bc9f. Read the comment docs.

srlanalytics commented 5 years ago

OK... should be good to merge now. The main thing here is putting forecasts of SA factors back in sa_we() with ahead = 0, (though maybe we should change it to forecast = 0 as forecast is probably more obvious?) as well as automatic log transformation or the option to specify 'log' or 'none'. This also deals with non-stationary series using loess() as discussed earlier. I'll keep working on things today... but probably good to merge soon to avoid more conflicts.