Data Mining

The course Data Mining provides an introduction to advanced data analysis techniques as a basis for analyzing business data and providing input for decision support systems. The course covers the following topics:

• Introduction to Data Mining: The course starts with a general overview of the most important types of data mining techniques as well as an overview of the major application areas of these techniques. We will discuss the main tasks within data mining projects, being data gathering and exploration, data pre-processing and transformation, pattern discovery, and evaluation/interpretation.
• Data Exploration: Before applying any data mining techniques, it is important to gain an initial understanding of the data to be mined. This includes calculating summarization statistics, visualizing the data, but also identifying data quality problems such as outliers or missing values. We will discuss dataset summarization and visualization techniques and will apply them to different data sets.
• Cluster Analysis: The goal of cluster analysis is, given a set of objects, to find groups of objects such that the objects in a group will be similar to one another and different from the objects in other groups. Application areas of cluster analysis are for example customer segmentation, market research, e-commerce, and image processing. We will discuss different partitional and hierarchical clustering techniques and apply these techniques to find groups of similar objects in different data sets.
• Classification: The goal of classification is to learn a function from training data which is able to assign previously unseen records to certain predefined classes (groups) as accurately as possible. Classification techniques are widely applied in different settings, including credit risk assessment, marketing, and fraud detection amongst others. We will cover several widely used classification techniques, including Decision Trees, k-NN, Rule Learning, Naïve Bayes, Artificial Neural Networks, and Support Vector Machines. We will also discuss how to evaluate and optimize the learned models.
• Association Analysis: Co-occurrence relationships between items are important with various application domains, such as marketing, supermarket shelf management, inventory management, and web usage analysis. We will cover different techniques for identifying co-occurrence patterns in data and will learn how to select the subset of relevant patterns from the usually very large pattern sets discovered by the algorithms.
• Text Mining: The term text miningrefers to the extraction of implicit, previously unknown and potentially useful information from a large amount of textual resources. Example applications of text mining include topical classification of news stories or web pages, email and news filtering, spam detection, sentiment analysis, and search term auto-completion. We will learn how to apply all the techniques that we have covered in the other sections not only to structured data but also to text and will experiment with different text clustering and classification tasks.