PokemonGoers / PredictPokemon-2

In this project we will apply machine learning to establish the TLN (Time, Location and Name - that is where pokemons will appear, at what date and time, and which Pokemon will it be) prediction in Pokemon Go.
Apache License 2.0
9 stars 3 forks source link

Classifier performance on the 50K data set #47

Closed goldbergtatyana closed 8 years ago

goldbergtatyana commented 8 years ago

What is your feature set, classifier and the performance on the 50K data set?

goldbergtatyana commented 8 years ago

@MatthiasBaur status?

MatthiasBaur commented 8 years ago

Status inc tonight.

sacdallago commented 8 years ago

😴 zombie issue

bensLine commented 8 years ago

true, running only attribute selection on the 50k file with 208 features in total takes already forever -_- feature set looks like this (there are 151 co-occurance features):

Results about the performance will follow with more details

goldbergtatyana commented 8 years ago

hm, feature selection is usually a really quick matter. it would be much helpful debugging your issues if you would provide your input file and the parameters you use for the tools.

MatthiasBaur commented 8 years ago

Performance was tested mainly on the 10k set. I will reference the results shortly.

MatthiasBaur commented 8 years ago

62