Master the art of predictive modeling with XGBoost and gain hands-on experience in building powerful regression, classification, and time series models using the XGBoost Python API
Key Features
- Get up and running with this quick-start guide to building a classifier using XGBoost
- Get an easy-to-follow, in-depth explanation of the XGBoost technical paper
- Leverage XGBoost for time series forecasting by using moving average, frequency, and window methods
- Purchase of the print or Kindle book includes a free PDF eBook
Book Description
XGBoost offers a powerful solution for regression and time series analysis, enabling you to build accurate and efficient predictive models. In this book, the authors draw on their combined experience of 40+ years in the semiconductor industry to help you harness the full potential of XGBoost, from understanding its core concepts to implementing real-world applications.
As you progress, you'll get to grips with the XGBoost algorithm, including its mathematical underpinnings and its advantages over other ensemble methods. You'll learn when to choose XGBoost over other predictive modeling techniques, and get hands-on guidance on implementing XGBoost using both the Python API and scikit-learn API. You'll also get to grips with essential techniques for time series data, including feature engineering, handling lag features, encoding techniques, and evaluating model performance. A unique aspect of this book is the chapter on model interpretability, where you'll use tools such as SHAP, LIME, ELI5, and Partial Dependence Plots (PDP) to understand your XGBoost models. Throughout the book, you’ll work through several hands-on exercises and real-world datasets.
By the end of this book, you'll not only be building accurate models but will also be able to deploy and maintain them effectively, ensuring your solutions deliver real-world impact.
What you will learn
- Build a strong, intuitive understanding of the XGBoost algorithm and its benefits
- Implement XGBoost using the Python API for practical applications
- Evaluate model performance using appropriate metrics
- Deploy XGBoost models into production environments
- Handle complex datasets and extract valuable insights
- Gain practical experience in feature engineering, feature selection, and categorical encoding
Who this book is for
This book is for data scientists, machine learning practitioners, analysts, and professionals interested in predictive modeling and time series analysis. Basic coding knowledge and familiarity with Python, GitHub, and other DevOps tools are required.
Table of Contents
- An Overview of Machine Learning, Classification, and Regression
- XGBoost Quick Start Guide with an Iris Data Case Study
- Demystifying the XGBoost Paper
- Adding On to the Quick Start – Switching Out the Dataset with a Housing Data Case Study
- Classification and Regression Trees, Ensembles, and Deep Learning Models – What's Best for Your Data?
- Data Cleaning, Imbalanced Data, and Other Data Problems
- Feature Engineering
- Encoding Techniques for Categorical Features
- Using XGBoost for Time Series Forecasting
- Model Interpretability, Explainability, and Feature Importance with XGBoost
- Metrics for Model Evaluations and Comparisons
- Managing a Feature Engineering Pipeline in Training and Inference
- Deploying Your XGBoost Model
Partha Pritam Deka is a data science leader with 15+ years of experience in semiconductor supply chain and manufacturing. As a senior staff engineer at Intel, he has led AI and machine learning teams, achieving significant cost savings and optimizations. He and his team developed a computer vision system that improved Intel's logistics, earning CSCMP Innovation Award finalist recognition. An active AI community member, Partha is a senior IEEE member and speaker at Intel's AI Everywhere conference. He also reviews for NeurIPS, contributing to AI and analytics in semiconductor manufacturing.
Joyce Weiner is a principal engineer with Intel Corporation. She has over 25 years of experience in the semiconductor industry, having worked in fabrication, assembly and testing, and design. Since the early 2000s, she has deployed applications that use machine learning. Joyce is a black belt in Lean Six Sigma and her area of technical expertise is the application of data science to improve efficiency. She has a BS in Physics from Rensselaer Polytechnic Institute and an MS in Optical Sciences from the University of Arizona.
