main-fsm.cpp File Reference

The main function of frequent itemsequence mining algorithms. More...

#include "common.hpp"
#include "io/input/transaction_reader/brBufferedTransactionReader.hpp"
#include "io/codec/decoder/df/CacheDFDecoder.hpp"
#include "util/StreamParser.hpp"
#include "util/SeqFrequentFilter.cpp"
#include "datastructures/maxvector.hpp"
#include "datastructures/trie/edgelist/OrderedEdgelist.hpp"
#include "apriori/bodon/Trie.hpp"
#include "apriori/SeqAprioriSelector.hpp"
#include <vector>
#include <iostream>
#include <string>

Include dependency graph for main-fsm.cpp:

Go to the source code of this file.

Functions

void init ()

void usage ()

commandline -- some utility methods for fim commandline tools.

int process_arguments (int argc, char *argv[], counter_t &min_supp, bool &isrel, double &relminsupp, unsigned int &maxsize)

This procedure processes the arguments.

int main (int argc, char *argv[])

Variables

std::string file_format

Detailed Description

The main function of frequent itemsequence mining algorithms.

Mining frequent sequence of item is a natural generalization of frequent itemset mining (FIM). Given a sequence of transactions, where each transaction is a sequence over an alphabet $I$ , we have to find the sequences that occure as a subsequence in at least $min\_supp$ number of transactions. A sequence $s=\langle i_1, i_2, \ldots, i_n\rangle$ is a subsequence of $s'=\langle i'_1, i'_2, \ldots, i'_m\rangle$ if there exist integers $1\le j_1 < j_2 < \cdots < j_n \le m$ , such that $i_1 = i'_{j_1}, i_2 = i'_{j_2}, \ldots, i_n = i'_{j_n}$ , i.e. we can get $s$ by deleting some items from $s'$ . For example $\langle e, a, a, b\rangle \prec \langle f, e, a, b, c, a, a, c, b\rangle$ because $i_1 = 2$, $i_2 = 3$, $i_4 = 6$, $i_4 = 9$ meet the requirements.

Currently this program only contains a trie-based Apriori implementation. For more information about this solution the reader is referred to the paper "A Trie-based APRIORI Implementation for Mining Frequent Item sequences" from Ferenc Bodon. It can be downloaded from the webpage of OSDM'05 workshop of ACM SIGKDD (http://www.cs.rpi.edu/~zaki/OSDM05/papers/p56-bodon.pdf).

Author:: Ferenc BODON

Date:: 2005-04-19

Definition in file main-fsm.cpp.

Function Documentation

void init ( )

Definition at line 52 of file main-fsm.cpp.
References file_format.

int main ( int argc,

char * argv[]

)

Definition at line 149 of file main-fsm.cpp.
References SeqFrequentFilter< IT_R >::findFrequentItems(), init(), process_arguments(), FileReprBase::READ, usage(), and FileReprBase::WRITE.

int process_arguments ( int argc,

char * argv[],

counter_t & min_supp,

bool & isrel,

double & relminsupp,

unsigned int & maxsize

)

This procedure processes the arguments.

Returns:

0, if no error is occured,

1, in case of an IO error,

2, in case of too few arguments,

3, if proper min_supp cannot be generated.

Definition at line 88 of file main-fsm.cpp.
References convert(), largest_itemsetsize, and usage().

void usage ( )

commandline -- some utility methods for fim commandline tools.

Definition at line 63 of file main-fsm.cpp.
References file_format.

Variable Documentation

std::string file_format

Definition at line 50 of file main-fsm.cpp.

Generated on Sun Sep 17 17:53:13 2006 for FIM environment by

1.4.4


Functions
void	init ()
void	usage ()
	commandline -- some utility methods for fim commandline tools.
int	process_arguments (int argc, char *argv[], counter_t &min_supp, bool &isrel, double &relminsupp, unsigned int &maxsize)
	This procedure processes the arguments.
int	main (int argc, char *argv[])
Variables
std::string	file_format