Looking at the first link from your post there is an in-depth overview of the different categories. Cat_#_level_1 denotes the top level category (12 in total) whereas Cat_#_level_2 denotes the second level category (up to 19 categories, depending on the level 1 category). E.g. an email with the labels Cat_3_level_1 and Cat_6_level_2 has the label california ...
In short theres no best proportion or rule of thumb.
It's highly dependent on how much data you have, its distribution in relation to the number of labels, and if samples are related to each other in anyway or are they completely independent of each other. You could look into k-fold training. Say you split your data into 5 sections each 20 percent of the ...
We have a Dataset
Dataset has Instances
Instances can be a Vector Or a Sequence.
If it is a vector, It has Features
Examples - Image, Tabular data
If it is a Sequence, It has Time-steps**
Each Time-step can have Features i.e. Uni/Multi-Variate
Examples - Video, Audio, Time-series data
**When we use text-data as a sequence, we call the words as "...