DL12 [PUBG] Detecting Outliers Detecting Outliers¶ 1. Head Shot¶ For many users, high head shot rate is almost impossible. Even professional online fps gamers hard to exceed 30%. headshotkills/kills = headshot_rate 2. Damagedealt¶3. kills¶4. killstreaks¶5. longestkill¶ As fas as I know, 1km kill is very hard to achieve. 6. rankpoints(elo-like ranking)¶7. revives¶8. roadkills¶9. swimdistance¶10. teamkills¶ For detecting abuser.. 2020. 6. 2. [PUBG] EDA In [1]: import os, time, gc import pandas as pd, numpy as np from tqdm import tqdm In [2]: os.listdir('input') Out[2]: ['sample_submission_V2.csv', 'test_V2.csv', 'train_V2.csv'] In [3]: %%time tr = pd.read_csv("input/train_V2.csv") te = pd.read_csv("input/test_V2.csv") Wall time: 13.1 s In [4]: tr.head() Out[4]: Id groupId matchId assists boosts damageDealt DBNOs.. 2020. 5. 29. [tabnet] beating tablet data with deep learning github.com/dreamquark-ai/tabnet dreamquark-ai/tabnet PyTorch implementation of TabNet paper. Contribute to dreamquark-ai/tabnet development by creating an account on GitHub. github.com www.youtube.com/watch?v=ysBaZO8YmX8&feature=youtu.be data-newbie.tistory.com/377 TABNET: ATTENTIVE INTERPRETABLE TABULAR LEARNING -1 https://github.com/google-research/google-research/tree/master/tabnet https://ar.. 2020. 5. 28. 이전 1 2 다음