Pokemon dataset proposals(and minor considerations) - AAU-Dat/P5-Nonlinear-Dimensionality-Reduction GitHub Wiki

Potential Pokemon datasets:

Exhibit nr 1:

https://www.kaggle.com/datasets/vishalsubbiah/pokemon-images-and-types?resource=download&select=pokemon.csv

This dataset contains one image for each pokemon + csv file with name and types.

Miscellaneous for exhibit 1: there is an example of someone who used CNN to classify pokemons with the intent of predicting whether the opkemon would be fire-type or water-type based on the image. Link here: https://medium.com/m2mtechconnect/classifying-pok%C3%A9mon-images-with-machine-learning-79b9bc07c080#4334

Exhibit nr 2:

https://www.kaggle.com/datasets/rounakbanik/pokemon?datasetId=2756&sortBy=voteCount

This csv file contains all the info regarding the pokemons. It does not have any images.

Miscellaneous for exhibit 2: there are other infos in the csv that could potentially be merged with other exhibits. . . an idea would be to classify pokemons' type based on height :DDD

Exhibit nr 3:

https://www.kaggle.com/datasets/lantian773030/pokemonclassification

This dataset contains 7k images of pokemons.

Miscellaneous for exhibit 3: there is an LDA done for the stats. Link here: https://www.kaggle.com/code/residentmario/linear-discriminant-analysis-with-pokemon-stats/notebook

One could train the model (in case we are talking about ANNs) with the 7k images + exhibit 1.