This document proposes a novel parallel algorithm for mining frequent sequences that uses static load balancing. It splits the set of frequent sequences into equivalence classes and estimates the processing time of each class using sampling. This approach distributes the classes among processors to improve performance over existing methods like GSP, which are slow and memory intensive. The key modules estimate sequence support within transactions and the relative size of equivalence classes to determine processing load. The goal is to achieve near-linear speedup using this probabilistic load balancing technique for parallel frequent sequence mining.

1. #13/ 19, 1st Floor, Municipal Colony, Kangayanellore Road, Gandhi Nagar, Vellore – 6.
Off: 0416-2247353 / 6066663 Mo: +91 9500218218
Website: www.shakastech.com, Email - id: shakastech@gmail.com, info@shakastech.com
Probabilistic Static Load-Balancing Of Parallel Mining Of Frequent Sequences
ABSTRACT
Frequent sequence mining is well known and well studied problem in data mining. The
output of the algorithm is used in many other areas like bioinformatics, chemistry, and market
basket analysis. Unfortunately, the frequent sequence mining is computationally quite expensive.
In this paper, we present a novel parallel algorithm for mining of frequent sequences based on a
static load-balancing. The static load-balancing is done by measuring the computational time
using a probabilistic algorithm. For reasonable size of instance, the algorithms achieve speedups
up to =3/4 P where P is the number of processors. In the experimental evaluation, we show that
our method performs significantly better than the current state-of-the-art methods. The presented
approach is very universal: it can be used for static load-balancing of other pattern mining
algorithms such as item set/tree/graph mining algorithms.
EXISTING SYSTEM
Frequent pattern mining is an important data mining technique with a wide variety of
mined patterns. The mined frequent patterns can be sets of items (itemsets), sequences, graphs,
trees, etc. The GSP algorithm is the first to solve the problem of frequent sequence mining. As
the frequent sequence mining is an extension of itemset mining, the GSP algorithm is an
extension of the Apriori algorithm. The Apriori and the GSP algorithms are breadth-first search
algorithms. The GSP algorithm suffers with similar problems as the Apriori algorithm: it is slow
and memory consuming.
Disadvantages of Existing System:
1. The frequent sequence mining is computationally quite expensive
2. Existing algorithms are slow and memory consuming methods

2. #13/ 19, 1st Floor, Municipal Colony, Kangayanellore Road, Gandhi Nagar, Vellore – 6.
Off: 0416-2247353 / 6066663 Mo: +91 9500218218
Website: www.shakastech.com, Email - id: shakastech@gmail.com, info@shakastech.com
PROPOSED ALGORITHM
In this paper we propose a novel parallel method that statically load-balance the
computation. That is: the set of all frequent sequences is first split into Prefix-Based Equivalence
Classes (PBECs), the relative execution time of each PBEC is estimated and finally the PBECs
are assigned to processors. The method estimates the processing time of one PBEC by the
sequential Prefixspan algorithm using sampling. In propose system, it is important to be aware
that the running time of the sequential algorithm scales with: 1) the database size; 2) the number
of frequent sequences; 3) the number of embeddings of a frequent sequence in database
transactions.
Advantages of Proposed System:
1. It is significantly better than the existing methods
2. We improve the estimate of the processing time of a single PBEC
MODULES
1. Estimation of Support Module
2. Estimation of Relative Size of PBEC Module
MODULE DESCRIPTION:
Estimation of Support:
In this module, we can estimate whether a support of sequence is a subsequence of a
transaction in a database or not.
Estimation of Relative Size of PBEC:
The relative size of a PBEC can be used as the estimate of the relative processing time of
the PBEC by a sequential algorithm. This estimate ignores some details of the sequential
algorithm. The relative size of PBEC might be smaller and controllable size.

3. #13/ 19, 1st Floor, Municipal Colony, Kangayanellore Road, Gandhi Nagar, Vellore – 6.
Off: 0416-2247353 / 6066663 Mo: +91 9500218218
Website: www.shakastech.com, Email - id: shakastech@gmail.com, info@shakastech.com
SYSTEM REQUIREMENTS
HARDWARE REQUIREMENTS:
 Processor - Pentium –IV
 Speed - 1.1 Ghz
 Ram - 256 Mb
 Hard Disk - 20 Gb
 Key Board - Standard Windows Keyboard
 Mouse - Two or Three Button Mouse
 Monitor - SVGA
SOFTWARE REQUIREMENTS:
 Operating System - Windows XP
 Coding Language - Java