new paper about a recommender system

In a previous post I had written about
"A Novel K-Means Based Clustering Algorithm for High Dimensional Data Sets"
and in other post there is the summary of recommender systems.

Here is a new and fresh paper:
"A New Method for Dimensionality Reduction using K-Means Clustering Algorithm for High Dimensional Data Set" International Journal of Computer Applications (0975 – 8887) Volume 13– No.7, January 2011

DO NOT USE those formulas from recommender systems for similarity calculation in online dating sites because if you do so, your Online Dating Site will reach "as low as" 3 to 4 persons high compatible per 1,000 persons screened, so in a 1,000,000 women database, any man will see as many as 3,000 to 4,000 women to contact (nearly at the same time), that means, a whole precision LESS than anyone could achieve by searching on one's own!

The answer to solve the above problem is: "How LIFEPROJECT METHOD calculates similarity"