Share Description
We go over important modelling issues like missing values, non-linear relationships, and model selection in depth. The bias-variance tradeoff and its role in prediction are also discussed. Finally, we examine various methods for evaluating a model, including performance metrics and assessing both internal and external validity. We also go over how to adapt a model to a specific environment.
Throughout the course, we use R to demonstrate the concepts introduced in the lectures. To follow the course, you do not need to instal R on your computer; you will be able to access R and all of the example datasets from within the Coursera environment. We do, however, make references to additional packages that you can use for specific types of analyses – please feel free to instal and use them on your computer.
In addition, each module may include practise quiz questions. In these, whether you provided the correct or incorrect answer, you will pass. You will learn the most by first considering the answers and then cross-referencing your answers with the correct answers and explanations provided.
Syllabus :
1. (A) Welcome to Leiden University
- Welcome to the course Predictive Analytics
- How to succeed in your online class?
(B) Prediction for prevention, diagnosis, and effectiveness
- Introduction
- Introduction to predictive analytics
- Predictive analytics in prevention
- Predictive analytics diagnosis
- Predictive analytics in intervention
- To conclude
2. Modeling Concepts
- Introduction
- Design issues
- Sample size
- Overfitting
- Bootstrapping
- To conclude
3. Model development
- Introduction
- Missing values
- Continuous predictors
- Model selection
- Model estimation
- To conclude
4. Model validation and updating
- Introduction
- Performance measures
- Validation approaches
- Updating approaches
- Predictive analytics for Aruba
- To conclude
