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.

Feature Engineering

Arithmetische Features

Datumsfeatures

Kategorische Features

Label-Encoding

One-Hot-Encoding

Target Encoding

Mean Target Encoding

K-fold-Cross-Validation

Über binäre Klassifikation hinaus

Fehlende Daten

Fehlende Daten finden

Fehlende Daten imputieren

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 ist die bekannteste Plattform für Data-Science-Wettbewerbe. An solchen Wettbewerben teilzunehmen ermöglicht dir, mit realen Datensätzen zu arbeiten, verschiedene Machine-Learning-Probleme zu erkunden, dich mit anderen zu messen und dabei unschätzbare Praxis zu sammeln. In diesem Kurs lernst du, wie du jede Data-Science-Competition angehst und strukturierst. Du wirst in der Lage sein, das richtige lokale Validierungsschema zu wählen und Overfitting zu vermeiden. Außerdem beherrschst du fortgeschrittenes Feature Engineering sowie Verfahren zum Ensembling von Modellen. All diese Techniken wendest du auf Datensätze aus Kaggle-Competitions an.

Extreme Gradient Boosting with XGBoost

Erfahren Sie, wie man Kaggle-Wettbewerbe gewinnt: Richtige Strategie und Validierung wählen.

Eine Kaggle-Competition in Python gewinnen

Dieser Kurs richtet sich an alle Data Scientists, die Kaggle-Competitions angehen und gewinnen wollen.

Wissenschaftler für maschinelles Lernen in Python

Fehlende Daten finden

Interaktive praktische Übung