This 50-minute video introduces the fundamental concepts of Data Mining, a powerful analytical technology. We start by introducing to you the process of data mining and the SQL Server Analysis Services (SSAS) Data Mining architecture, using the Multidimensional and Data Mining Mode mode of SSAS. We introduce data mining tools, starting with Excel with the free Data Mining Add-Ins for Office, and focusing on SQL Server Data Tools (SSDT), which are well suited to longer-duration analytical mining projects.
The most fundamental concept in data mining is that of Cases, which represent the entities you wish to analyse, such as customers, products, or events. The simplest form of a case is just a flat, denormalized row of data. We briefly explain other formats of cases, too: Customer Signatures, which contains as-of validity dates, and Nested Cases, on which we focus towards the end of this tutorial, when you can also see a demo comparing the use of Decision Trees with, and without, nesting.
You will also learn about the concepts of Mining Structures, used to describe Cases, Mining Models, and Mining Algorithms. We briefly introduce 9 of the Microsoft data mining algorithms: Naïve Bayes, Clustering, Decision Trees, Association Rules, Sequence Clustering, Neural Networks, Logistics Regression, Linear Regression, and Time Series. They will be explained in more detail in other modules of this series.
The remainder of this module discuses Column Data Types (especially Text, Long, and Double), and Column Content Types, focussing on the differences between Continuous, Discrete, and Discretized data. You will hear about different approaches to automatic Discretization, including Equal Areas, Clusters, and Thresholds technique, and about assisting the algorithms by hinting the statistical distribution of data in a column, such as Normal, LogNormal, or Uniform.
To help you learn, there are 5 demos in this module, which you can follow using your own datasets, Adventure Works from GitHub, or by downloading our educational dataset, HappyCars, available when you purchase access to this course.
