Machine Learning for Data Science
Important information about the course:
Provided at this link.
Final Q&As (via webex)
Q&A 1: Tuesday Jan 31 from 16:15 to 17:45
https://fu-berlin.webex.com/fu-berlin/j.php?MTID=m963d4ddd4c0941df221195f4d6091236
Q&A 2: Thursday Feb 2 from 12:15 to 13:45
https://fu-berlin.webex.com/fu-berlin/j.php?MTID=ma8eb27f16c4520bc9887711719e6a9d8
Q&A 3: Tuesday Feb 7 from 16:15 to 17:45
https://fu-berlin.webex.com/fu-berlin/j.php?MTID=ma2dd3d6414904ddfbd8df1a8e5d40a7a
Course content:
The course provides an overview of machine learning (ML) methods and algorithms for data science. It consists of two parts:
The first part of the course discusses the emergence of big data, presents common ML-based visualization techniques for large datasets, and introduces basic ML methods such as principal component analysis, k-means clustering, linear regression, and linear discriminant.
The second part of the course introduces advanced ML techniques (probabilistic models, kernels, and deep learning) which address the uncertainty of the model and the nonlinearity of real-world data. It also covers the question of reproducibility and explainability of ML models.
By the end of the course, the student has become familiar with common ML models for data science, has understood their theoretical and algorithmic foundations, and has gained experience on how to use these models adequately in practical situations.
List of topics:
- Big data and data science
- ML-based data visualization
- Principal components, anomalies, and clustering
- Correlation analysis, linear regression, and linear discriminant
- Reproducibility
- Sequential models
- Probabilistic models
- Kernel methods
- Deep learning
- Explainable AI