# Application of Association rule analysis to study the evolution of halos in Cosmological N-Body simulations

## TL;DR

This paper introduces a novel method using association rule analysis from data mining to construct halo merger trees in cosmological N-body simulations, providing new insights into halo evolution and substructure.

## Contribution

The study applies association rule analysis to build halo merger trees from simulation data, offering an alternative to traditional methods and enabling extraction of halo substructure and splitting information.

## Key findings

- Successfully constructed merger trees using association rules
- Compared halo accretion histories with traditional merger tree methods
- Demonstrated potential for analyzing halo substructure and splitting

## Abstract

Merger trees track the evolution of halos across multiple snapshots. They assign for halos of a particular snapshot, the set of halos from previous snapshots they possibly originated from. In this work, Association rule analysis a well known technique from data mining has been used to build halo merger trees. Association rule analysis tries to find associations between different halos(in same as well as in snapshots) using the particle IDs of the particles which the halos are made of. Associations are expressed in the form of association rules. Merger trees are one of the several useful results one can obtain from the output of association rule analysis. Other results including halo substructure and halo splitting can also be extracted. Each type of output to be extracted from the association rule analysis output correspond to a pattern in association rules. Merger trees were formed and tested using the above technique. Dark matter simulations were run using Gadget-2 for 128^3 particles. Halos were extracted from the simulation snapshots using Amiga Halo Finder. Halo accretion history was plotted and compared against those formed using AHF merger tree builder.