Does Mathematica support collaborative filtering and recommendation systems such as can be found in Apache Mahout?
I could not find any resources about advanced data mining techniques, beyond clustering and NN algorithms. Does anyone know any Mathematica libraries that support advanced data mining?
I'm building a specific type of a recommendation engine.
- each product can belong to many categories
- categories form a forest
- products are organized in a graph to show the relationship that consuming an entry level product X will make the more advanced product Y more suitable for the consumer
- for each consumer we have a timeseries of choosen products and the objective and subjective satisfaction levels
- ranked list of products per category that would maximize the objective satisfaction level while still in a threshold of the subjective satisfaction
- Rank products according to subjective satisfaction but occuring in preferable circumstances (if you've found the right patterns of when to introduce a product use product rating when introduced in a similar pattern)