Simple LSTM-based lottery forecast deep learning model.
The dataset is collected from official korea lottery Web: https://dhlottery.co.kr/common.do?method=main
- tensorflow >= 2.0
- keras
- sklearn
- numpy
- pandas
run main.py
script
or in your prompt, type
python main.py --data_dir 'dataset/lottery_history.csv' --mode 'eval_model' --trial 5 --training_lengt 0.90
dataset version is on the date: [21.04.23]
you can update dataset from the Web: https://dhlottery.co.kr/common.do?method=main
Arguments:
training_length : there are ~980 lottery cases in total. you can decide to what extent to use as for training length. (e.g. 0.5 training_lengh uses 485 lottery cases are used for training and infer 486 th as a test set.)