Thanks to Albert Strusberg, David Burg, Perrin Meyer, and Jason Walker-Yung we enjoy Loglet Lab 5.1.
What’s new:
Advanced the user interface:
Added graphing option form processing in the back-end.
Added scaling and normalizing function to process the data accordingly in the back-end.
Added the forecast for each line in the composite and components plot.
Completed corresponding Javascript (JS) coding to process the new data in the front-end graphs.
Adapted current database structure to save new parameters to library entries.
Adapted export/import logic to take into consideration all new parameters from the advanced UI.
Updated production database in the AWS server to the new database model.
Implemented the Logistic Substitution (LogSub) algorithm:
Created a multi-step algorithm in several modules to handle Logsub dataset creation based on the current input spreadsheet style form.
Adjusted the front-end logic to process the Logsub plotly JS data accordingly when using the Logsub model.
Adjusted the library logic to handle loading and saving Logsub data.
Adjusted the import/export logic to also save Logsub data when required.
Created new type of graph using Plotly JS for stack/market share data.
New features and miscellaneous bug fixes:
Added statistics to the advanced UI table. Statistics now include several parameters: RMS, MAD, MAPE and AIC.
Fixed bug when saving certain kind of datasets to the library.
Fixed X ranges calculation for the composite graph.
Fixed several import/export bugs for certain kind of datasets.
Fixed logic to handle certain kind of datasets when some of the rows contains NAN or zeros values.
Fixed the guessing initial parameters logic for the Logsub model.
All datasets transferred to from LL4 to LL5’s library
Documentation
Fixes and typos corrected
Additions according to the changes for v5.1
AWS:
LL4 has been removsed from AWS. Code is still available on Github.
LL5 – live long and prosper!