I am taking a class in data mining and learned a few things I wanted to share. As you might know, data mining is a process of finding previously unknown, non-trivial patterns from high volumes of data. It is highly used in industry where actionable results are required to make business decision. I will start with two different process of data mining.

Unsupervised Data Mining: There is no target variable. One of the examples is Association Analysis (AA), which is also called Market Basket Analysis. Association analysis basically tells you what things occur together.

Supervised Data Mining: In this technique, target variable is present. For instance: neural network and decision tree. Decision tree has a target variable set, and it helps us understand the path is takes to reach to the leaf from the root note.

Collaborative Filtering, recommendation systems, intelligent systems, etc are very fascinating to me. During my free time, I read about these systems and look for ways how I can utilize this to build great products. I have been following recommendation systems build by Amazon and other companies for a long time. Recently, I had to work on one of these projects where I had to generate item-to-item recommendation system based off of statistics collected from IIS logs. As this is still a proprietary product, I cannot discuss codes or process used to generate this. But the application was heavily influenced by Slope One algorithm and implemented via T-SQL and SSIS. If you are looking forward to build something of this nature, http://en.wikipedia.org/wiki/Slope_One and the corresponding links might be a good place to start with.