numpy, pandas, sklearn, gensim, xgboost, ripser, persim

3000 texts created by human and 3000 texts created by AI

generated column: 0 is for human and 1 is for AI

data is already processed, all texts are lowercase and have no punctuation

for each text:

1. take 120 most frequent words
2. vectorize all of them -> this will be our point cloud
3. calculate persistent diagram on obtain cloud
4. calculate persistence image of this diagram
5. train models with persistence image as feature

score on test: 0.85

only 4500 texts used for training

original dataset: https://www.kaggle.com/datasets/shanegerami/ai-vs-human-text/data

persim paper: https://arxiv.org/abs/1507.06217