You're writing a PPM compressor. It's different from mixing bit predictions. Read about ppmd in my book. You want to use the longest possible context and estimate the probability of a novel value (zero frequency) that you divide among the next smaller context. That can be learned.
Also get on encode.su. Lots of compression developers there can give you advice. On Fri, Feb 28, 2020, 11:52 PM <[email protected]> wrote: > matt, > let's use 3 terms 'total unique symbols', 'total counts', and 'individual > counts. > > Ex. we have 5 prediction sets to mix, and set_1 is: > a 5, b 2, c 3, d 8, e 1 > > total unique symbols=5 > total counts=19 > individual counts=5, 2, 3, 8, 1 > > I know, I know, the amount we mix this set will be determined by how many > 'total counts' (19), but, what else determines its weight? > > *Artificial General Intelligence List <https://agi.topicbox.com/latest>* > / AGI / see discussions <https://agi.topicbox.com/groups/agi> + > participants <https://agi.topicbox.com/groups/agi/members> + delivery > options <https://agi.topicbox.com/groups/agi/subscription> Permalink > <https://agi.topicbox.com/groups/agi/Tcfc4df5e57c62b43-M42db3c488177bc876b4ba8c0> > ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Tcfc4df5e57c62b43-M7f89fa25cef6657504909680 Delivery options: https://agi.topicbox.com/groups/agi/subscription
