# Efficient Analysis of Pattern and Association Rule Mining Approaches

**Authors:** Thabet Slimani, Amor Lazzez  
**arXiv:** [1402.2892](https://arxiv.org/abs/1402.2892)  
**Date:** 2014-02-12  
**Summary page:** https://www.summarizepaper.com/en/paper/1402.2892v1/

## Key points

- Discovering patterns from a given data source is crucial in data mining
- Most recognized data mining tasks include frequent itemset mining, sequential pattern mining, structured pattern mining, and correlation mining
- Numerous efficient algorithms have been proposed for accomplishing these tasks
- Frequent pattern mining has received considerable attention due to its practical applications in various fields such as market basket analysis, web usage analysis, bioinformatics, etc.
- Several efficient algorithms have been developed for frequent itemset discovery using different strategies such as Apriori-based methods, FP-growth algorithm, Eclat algorithm, etc.
- Sequential pattern mining involves discovering frequently occurring subsequences or episodes within sequences or event logs. Several approaches have been proposed for this task such as GSP (Generalized Sequential Pattern), SPADE (Sequential Pattern Discovery using Equivalence classes), PrefixSpan (Prefix-Projected Sequential Pattern), etc.
- Structured pattern mining involves discovering patterns that satisfy certain structural constraints on attributes or relationships between them. Popular algorithms used for structured pattern mining include Subdue (Substructure Discovery), gSpan (Graph-Based Substructure Pattern Mining), etc.
- Correlation mining involves discovering interesting correlations between different attributes in a dataset. Several approaches have been proposed for this task such as Pearson's correlation coefficient, Spearman's rank correlation coefficient, Kendall's tau-b correlation coefficient, etc.
- Efficient algorithms are important for data mining tasks and promising research directions exist in this field.

## Summary

In the field of data mining, the process of discovering patterns from a given data source is crucial. Among the most recognized data mining tasks are frequent itemset mining, sequential pattern mining, structured pattern mining, and correlation mining. To accomplish these tasks, numerous efficient algorithms have been proposed in literature. Frequent pattern mining has been a focused topic in data mining research with significant progress made over time. This progress ranges from performant algorithms for frequent itemset mining in transaction databases to complex algorithms such as sequential pattern mining and structured pattern mining. Association Rule Mining (ARM) is one of the most current data mining techniques designed to group objects together from large databases aiming to extract interesting correlations and relations among huge amounts of data. They also discuss some promising research directions while including a comparative study between the performance of various approaches. The authors highlight that frequent pattern mining has received considerable attention due to its practical applications in various fields such as market basket analysis, web usage analysis, bioinformatics, etc. The article emphasizes that several efficient algorithms have been developed for frequent itemset discovery using different strategies such as Apriori-based methods, FP-growth algorithm, Eclat algorithm, etc. Moreover, the authors discuss sequential pattern mining which involves discovering frequently occurring subsequences or episodes within sequences or event logs. They mention that several approaches have been proposed for this task such as GSP (Generalized Sequential Pattern), SPADE (Sequential Pattern Discovery using Equivalence classes), PrefixSpan (Prefix-Projected Sequential Pattern), etc. Structured pattern mining is another area where significant progress has been made recently. It involves discovering patterns that satisfy certain structural constraints on attributes or relationships between them. The authors mention some popular algorithms used for structured pattern mining such as Subdue (Substructure Discovery), gSpan (Graph-Based Substructure Pattern Mining), etc. Finally, the article discusses correlation mining which involves discovering interesting correlations between different attributes in a dataset. The authors mention that several approaches have been proposed for this task such as Pearson's correlation coefficient, Spearman's rank correlation coefficient, Kendall's tau-b correlation coefficient, etc. In conclusion, Thabet Slimani and Amor Lazzez provide an insightful analysis of the current status of frequent pattern mining and its related techniques. They highlight the importance of efficient algorithms for data mining tasks and discuss promising research directions in this field.