A recommendation system for blogs: Setting up the prerequisites (part 1)

The goal of data science is typically described as creating value from Big Data. However, data science should also meet a second goal, that is, avoiding an information overload. One particular type of projects that really meet these two goals are recommendation engines. Online stores such as Amazon but also streaming services such as Netflix suffer from information overload. Customers can easily get lost in their large variety (millions) of products or movies. Recommendation engines help users narrow down the large variety by presenting possible options.

--

For example, a Netflix user who watched the movie Frozen gets similar children movies from Pixar as a recommendation to watch.

--

Roughly speaking we can divide recommendation engines in two different types: collaborative filtering and content-based.  

--

Collaborative filtering

Finding patterns among multiple readers. If several readers are interested in a particular set of articles it is very likely that a reader who starts reading one of these articles is also interested in the other articles from this set. Therefore, based upon the reading behavior of other users, suggestions are made to similar users. 

--

Content-based recommendation engines are different as they base their recommendations on the properties of the product.

If a user is reading an article containing the words 'Google Analytics' and 'Tag Manager', chances are that this user also likes reading other articles containing these words. Therefore, a content-based recommendation engine will recommend articles containing these words.  

https://www.themarketingtechnologist.co/building-a-recommendation-engine-for-geek-setting-up-the-prerequisites-13/