Search code examples
hadoopmahoutrecommendation-enginemahout-recommender

Coding to find Z score in Apache Mahout and compute similarity


I am new to apache mahout. I have managed to use it for pearson corelation and cosine vector but i need to normalize data and use Z Score to calculate similarity. I am unable to find methods in mahout which allow to do so. The mahout wiki also doesn't demonstrate the use of normalization of data and use for calculating similarity. I would be very thankful if someone can help me out with the code for the same.


Solution

  • These questions are better answered on the mahout user mailing list.

    In any case, it would be nice to understand what you are trying to do on a larger scale. It sounds like you might be trying to build a recommendation engine. If so, Pearson correlation is generally a really bad way to do that.

    It is much better to use Mahout to compute indicator behaviors and then use a search engine such as Solr or ElasticSearch to deploy the recommendation function.

    We described how to do this in the O'Reilly small book that you can get from:

    https://www.mapr.com/practical-machine-learning