README

Welcome to the experimental Machine Learning Theatre Days!

This is still a work-in-progress.

How to Use:

1. install Anaconda & Jupyter Notebook
2. CD to New, as the old is there for archieving purposes only
3. install requirements
4. run MLTD_v5
5. ????
6. Profit

How it works

This model works by first separating anniv data from regular events, then using the anniv data to determine each idol's popularity values to be used in the model. Of course there will be outliers as the in game events does have too. Some features are difficult to quantify such as song quality, seiyuu appearance, gacha cycles and event schedules themselves. Take the result of this model with a grain of salt.