Medical Risk Prediction Models: With Ties to Machine Learning

Medical Risk Prediction Models: With Ties to Machine Learning is a practical guide for clinicians, epidemiologists, and professional statisticians to create or assess statistical prediction models based on data. It covers the mathematical details of model creation and evaluation, including discrimination, calibration, and predictive performance with censored data and competing risks, with R-code and illustrative examples. It also provides benchmarks for interpretation of prediction performance and comparison and combination of rival modeling strategies via cross-validation.

Format: Hardback
Length: 290 pages
Publication date: 01 February 2021
Publisher: Taylor & Francis Ltd

The book is divided into three parts: Part I: Introduction to Statistical Prediction Models, Part II: Mathematical Details of Statistical Prediction Models, and Part III: Applications of Statistical Prediction Models.
Medical Risk Prediction Models: With Ties to Machine Learning is a comprehensive guide for clinicians, epidemiologists, and professional statisticians who are involved in developing or evaluating statistical prediction models based on data. The book focuses on the individualized probability of a medical event occurring within a specific time horizon. Gerds and Kattan provide a clear and pedagogical explanation of the mathematical details involved in creating and evaluating statistical prediction models, without resorting to mathematical notation. This book is an invaluable resource for anyone who is unsure about whether a Cox regression model outperforms a random survival forest.

Key Features:

1. Comprehensive Coverage: The book covers all the essential aspects of statistical prediction models, including discrimination, calibration, predictive performance with censored data and competing risks, R-code and illustrative examples, interpretation of prediction performance via benchmarks, and comparison and combination of rival modeling strategies via cross-validation.

2. Easy-to-Follow Format: The book is organized into three parts: Part I: Introduction to Statistical Prediction Models, Part II: Mathematical Details of Statistical Prediction Models, and Part III: Applications of Statistical Prediction Models. Each part is designed to be accessible and easy to follow, making it suitable for beginners and experienced practitioners alike.

3. Real-World Examples: The book includes numerous real-world examples to illustrate the concepts and techniques discussed. These examples range from clinical trials to epidemiological studies, and provide valuable insights into the practical applications of statistical prediction models.

4. R-Code and Illustrative Examples: The book includes R-code and illustrative examples that allow readers to implement the models discussed in the book. This makes it easy for readers to apply the models to their own data and gain hands-on experience.

5. Interpretation of Prediction Performance: The book provides a detailed explanation of how to interpret the prediction performance of statistical prediction models. This includes benchmarks, such as the area under the receiver operating characteristic curve (AUC), and other metrics that can help practitioners make informed decisions about their models.

6. Comparison and Combination of Rival Modeling Strategies: The book discusses the comparison and combination of rival modeling strategies via cross-validation. This allows practitioners to identify the best model for their specific data and to improve the performance of their models.

Part I: Introduction to Statistical Prediction Models

In Part I, the book provides an overview of statistical prediction models. It covers the basic concepts of statistical prediction, including regression analysis, logistic regression, and survival analysis. The book also discusses the importance of data quality and the need for appropriate data preprocessing.

Part II: Mathematical Details of Statistical Prediction Models

In Part II, the book delves into the mathematical details of statistical prediction models. It covers topics such as model selection, model fitting, and model evaluation. The book also discusses the use of cross-validation to assess the performance of statistical prediction models.

Part III: Applications of Statistical Prediction Models

In Part III, the book provides applications of statistical prediction models in various fields. It covers topics such as clinical trials, epidemiological studies, and public health. The book also discusses the ethical considerations involved in the use of statistical prediction models.

Conclusion:

Medical Risk Prediction Models: With Ties to Machine Learning is a valuable resource for clinicians, epidemiologists, and professional statisticians who are involved in developing or evaluating statistical prediction models based on data. The book provides comprehensive coverage of the mathematical details involved in creating and evaluating statistical prediction models, real-world examples, R-code and illustrative examples, interpretation of prediction performance, and comparison and combination of rival modeling strategies. Whether you are a beginner or an experienced practitioner, this book will help you gain a deeper understanding of statistical prediction models and improve your ability to make informed decisions about your data.

Weight: 606g
Dimension: 160 x 242 x 27 (mm)
ISBN-13: 9781138384477