sayakpaul / benchmarking-and-mli-experiments-on-the-adult-dataset

Contains benchmarking and interpretability experiments on the Adult dataset using several libraries

My name is Sayak Paul! 👾

• 🔭 I work on 🧨 diffusion models at Hugging Face 🤗
• 🌱 I’m interested in the area of representation learning.
• 👯 I’m always open to meaningful collaborations.
• ⚡ Fun fact: I love watching crime and action thrillers (The Silence of the Lambs being an all-time favorite one).
• 🙃 Recepient of the Google Open Source Peer Bonus Award (2020, 2021, and 2022). Also received the TensorFlow Top Contributor Award 2021.
• 📫 More details - sayak.dev.

Following are some of my favorite repositories that I have contributed to and/or contribute to.

benchmarking-and-mli-experiments-on-the-adult-dataset's Issues

Benchmark Fix for RandomForest

In the benchmark file

We use the LabelEncoder on our training and testing data separately, which will cause index mismatch.

from sklearn.preprocessing import LabelEncoder

train_preprocessed = train.apply(LabelEncoder().fit_transform)
test_preprocessed = test.apply(LabelEncoder().fit_transform)

We should change it with OrdinalEncoder

from sklearn.preprocessing import OrdinalEncoder
oe = OrdinalEncoder(handle_unknown='use_encoded_value', unknown_value=-1)

train_preprocessed = oe.fit_transform(train.fillna("NaN"))
test_preprocessed = oe.transform(test.fillna("NaN"))
train_preprocessed = pd.DataFrame(train_preprocessed, columns=train.columns)
test_preprocessed = pd.DataFrame(test_preprocessed, columns=train.columns)

With this transformation, I am able to get: 0.8526 accuracy for Random Forest.