Skip to content
This repository has been archived by the owner on Dec 23, 2020. It is now read-only.

Latest commit

 

History

History
71 lines (55 loc) · 2.52 KB

SSD_Def.md

File metadata and controls

71 lines (55 loc) · 2.52 KB
Raw

Subspaces Datasheet File (SSD)

Introduction

This file describes the layout of SSDs. It was created with the visualization tool Black Sheep Vision (BSV).

A SSD contains collected informations computed by a data mining algorithm. That means a SSD contains all subspaces where potentially outliers are detected by a data mining algorithm and a value, called outlierness value, which says if the object a is potential outlier in subspace b or not.

So a SSD is composed of two parts. In the first part are all subspaces are indexed where are potential outliers got detected. And in the second part are listed all outlierness values of each object in the corresponded subspaces.

Format

SSD are an extended csv format which is composed by two parts.

In part one all subspaces detected by a data mining algorithm will be indexed. A valid index entry in part one consits of the keyword "subspace" followed by its index followed by the keyword "dimension" followed by its dimension and followed by a base consits of list with all features, seperated with a comma, spanning this subspaces bracketed with squared braces.

Whitespaces in the base of the indexed subspaces are optional.

After that part two will be introduced with a "@data" tag followed by all outlierness values, computed by the same data mining algorithm which found the subspaces indexed in part one. A valid entry in part two consits of the "object id", the "subspace id" and the outlierness value of the object with "object id" in subspace "subspace id". All values have to be seperated by a valid delimiter of the csv format, that means valid you can use ',', ';', ':', ' ' or a tab as delimiter.

If you're using ',', ';' or ':' as delimiter all whitespaces are optional.

General SSD Layout

Here we see the general structure of a valid SSD.

subspace 0 dimension = x [F1, F2, ..., Fx]
  .		.
  .		.
  .		.
  .		.
subspace n dimension = y [P1, P2, ..., Py]

@data
obj_0; subspace_0; <<Detected Outlierness of obj_0 in subpsace_0>>
obj_0; subspace_1; <<Detected Outlierness of obj_0 in subpsace_1>>
  .	.		.
  .	.		.
  .	.		.
obj_0; subspace_n; <<Detected Outlierness of obj_0 in subpsace_n>>
  .	.		.
  .	.		.
  .	.		.
obj_i; subspace_0; <<Detected Outlierness of obj_i in subpsace_0>>
obj_i; subspace_1; <<Detected Outlierness of obj_i in subpsace_1>>
  .	.		.
  .	.		.
  .	.		.
obj_i; subspace_n; <<Detected Outlierness of obj_i in subpsace_n>>