Parallel approach on distance based outlier detection on streaming data
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
project
src/main/scala
README
build.sbt

README

A parallel approach using the parallel, advanced, slicing and pmcod algorithms with the grid and metric partitioning for distance based outlier detection on streams.

Parameters:
--input <Input path of data>
--treeInput (Optional) <Input path for the creation of the tree in metric partitioning>
--window <window size>
--slide <slide size>
--dataset <dataset name for grid partitioning (works on stock, tao, fc, gauss)>
--k <number of neighbors>
--range <range for distance>
--algorithm <parallel,advanced,advanced_vp,slicing,pmcod>
--part (Optional) <Partitioning type - grid,metric>
--VPcount (Optional) <Number of elements for tree creation in metric partitioning>
--parallelism <Parallelism degree for Flink>