In this first chapter, you will get exposure to the Kaggle competition process. You will train a model and prepare a csv file ready for submission. You will learn the difference between Public and Private test splits, and how to prevent overfitting.

Competitions overview

Explore train data

Explore test data

Prepare your first submission

Determine a problem type

Train a simple model

Prepare a submission

Public vs Private leaderboard

What model is overfitting?

Train XGBoost models

Explore overfitting XGBoost

Kaggle competitions process

Now that you know the basics of Kaggle competitions, you will learn how to study the specific problem at hand. You will practice EDA and get to establish correct local validation strategies. You will also learn about data leakage.

Understand the problem

Understand the problem type

Define a competition metric

Initial EDA

EDA statistics

EDA plots I

EDA plots II

Local validation

K-fold cross-validation

Stratified K-fold

Validation usage

Time K-fold

Overall validation score

Dive into the Competition

You will now get exposure to different types of features. You will modify existing features and create new ones. Also, you will treat the missing data accordingly.

Rekayasa fitur

Fitur aritmetika

Fitur tanggal

Fitur kategorikal

Pengodean label

One-Hot encoding

Target encoding

Pengodean target rata-rata

Melampaui klasifikasi biner

Data hilang

Temukan data hilang

Imputasi data hilang

Feature Engineering

Time to bring everything together and build some models! In this last chapter, you will build a base model before tuning some hyperparameters and improving your results with ensembles. You will then get some final tips and tricks to help you compete more efficiently.

Baseline model

Replicate validation score

Baseline based on the date

Baseline based on the gradient boosting

Hyperparameter tuning

Grid search

2D grid search

Model ensembling

Model blending

Model stacking I

Model stacking II

Final tips

Testing Kaggle forum ideas

Select final submissions

Final thoughts

Modeling

Demand forecasting (train)

Demand forecasting (test)

House prices (train)

House prices (test)

Taxi rides (train)

Taxi rides (test)

Kaggle adalah platform paling terkenal untuk kompetisi Data Science. Berpartisipasi dalam kompetisi semacam ini memungkinkan Anda bekerja dengan himpunan data dunia nyata, mengeksplorasi berbagai permasalahan Machine Learning, bersaing dengan peserta lain, dan pada akhirnya memperoleh pengalaman praktik yang sangat berharga. Dalam kursus ini, Anda akan mempelajari cara mendekati dan menyusun kompetisi Data Science apa pun. Anda akan mampu memilih skema validasi lokal yang tepat dan menghindari overfitting. Selain itu, Anda akan menguasai rekayasa fitur tingkat lanjut beserta pendekatan penyusunan ansambel model. Semua teknik ini akan dipraktikkan pada himpunan data kompetisi Kaggle.

Extreme Gradient Boosting with XGBoost

Pelajari cara menang di kompetisi Kaggle, dari pendekatan tepat hingga skema validasi yang sesuai.

Fitur kategorikal

Create Your Free Account