Predictive analytics: data-based forecasts with AI and prediction models

In the first learning unit, discover how you can make well-founded decisions using time series analyses and changing perspectives. The concepts of predictive analytics are explained using illustrative examples.

In the second webinar, you will put what you have learned into practice and learn how to analyze and interpret time series using Python code. You will learn how to break down time series into components such as trend, seasonality and noise to gain a better understanding of the underlying patterns.

In the first practical phase, you will have the opportunity to apply the knowledge acquired in the chapter and the Python classes developed in the webinar to specific tasks yourself. You will use Jupyter notebooks and datasets with which you can reproduce and experiment with the exercises. 

Dive into the world of statistics, regression models and optimal decision making, and learn to look into the future with statistics. You'll learn all about the basics of regression to identify relationships between variables and derive predictions, as well as the balance between overfitting and better outcomes. You will then learn about multivariate regression, which uses multiple independent variables, and the challenges of creating these models. The chapter concludes with gradient methods for optimizing regression models and shows you how to achieve the optimal parameters to make accurate predictions.

In the third webinar, you will put your newly acquired knowledge into practice. Together with the trainers , you will develop a new regression class with which you can make precise statistical predictions for your time series objects.

In this practical phase, you will have the opportunity to apply the knowledge acquired in the previous chapters and the Python classes developed in the webinar directly to specific tasks. Jupyter notebooks and provided data sets will be used.

In this self-study unit, you will take an in-depth look at neural networks and their building blocks. You will learn how AI models work and the advantages and disadvantages of their respective building blocks. Learn all about:

• Concepts and functions of neural networks
• Activation functions in neural networks
• Introduction to deep neural networks and their potential
• Introduction to deep learning and the crucial role of the chain rule
• Gradients and function chains
• Disappearing and exploding gradients

In this webinar, you will implement neural networks in Python using Pytorch. You will pour the concepts you have learned into code and design your own deep neural networks in just a few steps.

In this practical phase, you will apply the Python classes developed in the webinar directly to specific tasks. Jupyter notebooks and provided data sets will be used.

In this learning unit, you will familiarize yourself with transformer models and find out why they are ideal for time series predictions. To begin with, you will reflect on the knowledge you have acquired so far about time series, prediction models and neural networks. You will link the content you have learned with core concepts of machine learning and artificial intelligence and learn how transformers work in detail. Finally, you will learn how transformers are used and why they are so successful.

In the last joint webinar, you and the trainers will implement Transformer models in Python using PyTorch. The concepts you learn, such as embeddings, positional coding and attention mechanisms, are implemented in flexible classes. This enables you to quickly and easily apply Transformer models to your own data.

As in every practical phase, you will have the opportunity to apply the knowledge acquired in the chapter and the Python classes developed in the webinar directly to specific tasks in Jupyter Notebooks and experiment with them.