Essentials of predictive analytics - incompany training

Predictive analysis using statistics, artificial intelligence and machine learning methods

This course efficiently introduces the essential data science skills needed to develop and use adequate prediction models for quantitative data-based decision making.

On the one hand, the principles of commonly used methods for regression, classification and detection of data clusters are discussed, with special attention to the consequences of big data aspects. On the other hand, practical examples will be used to illustrate how models can be validated, compared and used.

Unique is that during the course participants gain experience in the visual programming of data workflows, with which model-fit, -validation and -comparison can be easily executed in practice.

Using prediction models for decision making

After completion of this course:

You have an overview of commonly used methods for predictive modeling from the areas of statistics, artificial intelligence and machine learning.
You can develop these models for standard situations independently, using software such as IBM Modeler, SAS-Enterprise Guide and Enterprise Miner or Orange to visually program data workflows.
You have gained practical experience with validating, interpreting and comparing alternative models and their use for decision support.

Intended for

Professionals who are involved in the analysis of quantitative data and the use of decision support systems, and managers who want to be able to assess and compare the quality of developed models or who control and steer these processes. The course is also suitable for lecturers at universities or colleges of higher education who want to be informed about developments in the field of data science, data mining and data analytics.

In consultation with the participants this course can be taught in Dutch or English.

Also take a look at the overview of all data analytics and statistics courses.
Curious about the background of course leader Koo Rijpkema and his vision on the importance of data in the world of technology? Read the interview!

Information

Language
The program can be taught in English on request.
Program
During the course, practical examples and assignments are discussed. Participants gain experience with the use of the freely available source software Orange for visually programming data analysis workflows and interactively validating and comparing models.

Topics treated:
- Introduction to course, software and case-study.
- Principles in predictive analytics:
  - Data pre-processing & data reduction. over-fitting and model tuning. Resampling methods & regularization.
- Essentials of prediction modeling:
  - Overview of commonly used prediction methods, such as linear regression, neural networks, regression trees and nearest neighbor algorithms.
  - Selection, validation and use of predictive models in practice.
- Essentials of classification:
  - Overview of commonly used classification methods, such as logistic regression, neural networks, classification trees, support vector machines and nearest neighbor algorithms.
  - Selection, validation and use of classification models in practice.
- Essentials of cluster and segmentation models:
  - Overview of commonly used cluster and segmentation methods, such as hierarchical, k-Means and Louvain clustering.
  - Selection, validation and use of cluster and segmentation models in practice.
- Epilogue: review, trends and next steps.