Scalable And Efficient Outlier Detection In Large Distributed Data Sets With Mixed-Type Attributes