I have made my submissions for the kaggle competition, i will run the

re trained the models on only tournament games and the ChalkSeedDiff

made a similar womens neural network

Merge branch 'main' of hf.co:Jensen-holm/Nigl

model with 10k epochs and 0.001 learning rate works much better than

Update README.md

Update README.md

working on training neural networks and evaluating performance against

adding the opponent chalk seed so that we can make baseline predictions

creating a neural network for the mens data, predicting the win column

working on building a neural network for the mens dataset

split up the mens and womens data and beginning building a neural

making new super detailed games dataframe where each row is a game, and

adding all sorts of diff data to the team aggregated stats

went back and fixed an issue in pre processing, realized that I was not

got rid of unnessessary tournament and regular season split attributes

removing csv in wrong plac

removing modeling and processing that used to be specific to mens bball,

merging the regular season stats and the tournament stats together and

did the same merge and aggregation for the detailed tournament data

working through aggregating the regular season data, deleted the data

columns that did not need to be aggregated were being aggregated. I

working on a new pre processing file with a much better approach to the

made the region column reflext the actual string that represents the

making new columns Seed_Num and Region_Num so that the model can train

modeling a given teams tournament score 5 number summary (minus standard

made a new model on 300k epochs to predict just score distributions for

built a neural network that was trained on 1 million epochs to predict

adding in the mean of statistics into the aggregated summary dataset.

adding gitignore

saving MDetailedGamesOvrStats.csv from mens_preprocessnig.ipynb

adding the points allowed column to the aggregated metrics

aggregating and summarizing team stats in mens_pre_procssing so that we

readme initial

moving the 5k epoch model to the hugging face repo

copying the src from the github repository

moving all downloaded csv data into data directory

uploading provided kaggle data

initial commit