Scribd Logo
Upload

Welcome to Scribd!

  • 40 views

    Survey On Mining Ars Using Genetic Algorithm

    Uploaded by

    Anonymous TxPyX8c
    The document compares 18 references that use genetic algorithms for association rule mining. It includes the parameters they use for selection, fitness evaluation, crossover and mutation methods. The selection methods include roulette wheel selection, tournament selection and niching. Fitness is based on measures like support, confidence, comprehensibility and accuracy. Crossover techniques include single point, uniform and multipoint. Mutation rates and positions are adapted in some. The algorithms are tested on datasets like nursery, mushroom, chess and abalone to evaluate number of rules mined and execution time.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    Survey On Mining Ars Using Genetic Algorithm

    Uploaded by

    Anonymous TxPyX8c
    40 views4 pages
    The document compares 18 references that use genetic algorithms for association rule mining. It includes the parameters they use for selection, fitness evaluation, crossover and mutation methods. The selection methods include roulette wheel selection, tournament selection and niching. Fitness is based on measures like support, confidence, comprehensibility and accuracy. Crossover techniques include single point, uniform and multipoint. Mutation rates and positions are adapted in some. The algorithms are tested on datasets like nursery, mushroom, chess and abalone to evaluate number of rules mined and execution time.

    Original Description:

    ewt

    Original Title

    Survey Table2

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document compares 18 references that use genetic algorithms for association rule mining. It includes the parameters they use for selection, fitness evaluation, crossover and mutation methods. The selection methods include roulette wheel selection, tournament selection and niching. Fitness is based on measures like support, confidence, comprehensibility and accuracy. Crossover techniques include single point, uniform and multipoint. Mutation rates and positions are adapted in some. The algorithms are tested on datasets like nursery, mushroom, chess and abalone to evaluate number of rules mined and execution time.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    40 views4 pages

    Survey On Mining Ars Using Genetic Algorithm

    Uploaded by

    Anonymous TxPyX8c
    The document compares 18 references that use genetic algorithms for association rule mining. It includes the parameters they use for selection, fitness evaluation, crossover and mutation methods. The selection methods include roulette wheel selection, tournament selection and niching. Fitness is based on measures like support, confidence, comprehensibility and accuracy. Crossover techniques include single point, uniform and multipoint. Mutation rates and positions are adapted in some. The algorithms are tested on datasets like nursery, mushroom, chess and abalone to evaluate number of rules mined and execution time.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    of 4

    Survey on Mining ARs Using Genetic Algorithm

    Table: Comparison of Parameters involved in Mining ARs using Genetic Algorithm

    Ref.
    No.
    Ref1

    Ref2

    Ref3

    Ref4

    Selection

    Fitness

    Niched Pareto Based


    on
    Selection
    comprehensibility
    and
    confidence
    factor
    Roulette
    Wheel
    Selection

    Rank based
    Pareto
    Dominance

    Niched Pareto Based


    on
    Selection
    comprehensibility
    and
    predictive
    accuracy
    Based on
    sustaining index,
    creditable index,
    Inclusive index.

    Crossover

    Mutation

    Normal

    Normal

    Accuracy Measure
    Confidence factor.

    Dataset

    Remarks

    Nursery

    Execution time compared

    Comprehensibility

    Zoo

    Predictive Accuracy

    Nursery

    Rule Interestingness

    Adult

    Insert , Remove operations


    control size of rules thereby
    influencing
    comprehensibility

    Comprehensibility

    Zoo.

    Predictive Accuracy

    Nursery

    Comprehension.

    Single point
    uniform
    crossover

    Normal

    Uniform
    Crossover

    Normal

    Repair operator used after


    crossover to create valid
    chromosomes
    Insert , Remove operations
    control size of rules

    Random
    crossover

    Random
    mutation

    Predictive Accuracy
    Simplicity of Rules

    Tic-tac toe

    Population initialization
    Dermatalogy based on entropy.
    Rule pruning done to
    Hepatitis
    remove irrelevant terms.
    Cleveleand

    Heart
    Disease.
    Ref5

    Tournament
    Selection

    Based on
    simplicity
    operator

    Symbiotic
    combination
    operator

    Normal

    Classification Rate

    CRX

    Execution time compared

    Iris
    Vote
    Wine
    KDD99

    Ref6

    Strength of
    implication

    Single
    crossover
    operator

    Bases
    position
    mutation

    Strength of
    implication, fitness

    Ref7

    Roulette
    Wheel
    Selection

    Based on
    completeness
    measure,
    confidence factor

    Random
    crossover

    Normal

    Based on minimum
    support and
    minimum confidence

    Ref8

    Random
    selection

    Based on support
    factor

    Order 1
    crossover

    Number of rules
    generated for given
    minimum confidence
    and support

    Ref9

    Selection
    based on
    fitness values

    Done by
    swapping
    data after
    going
    through all
    chromosome
    in population
    Mutation
    applied to all
    genes both at
    first and
    second level

    Based on
    confidence

    Crossover
    applies only
    to first level
    genes

    Predictive accuracy
    Sensitivity test

    Car test
    result
    Lens

    The frequent itemsets with


    min support are found by
    Apriori method and
    algorithm applied to this
    output.
    SPECT heart Minimum support and
    Solar flare
    minimum confidence are
    Nursery
    not considered
    Monks
    problems
    Balance
    Scale
    Stock
    Chromosome representation
    trading data is multiple level phenotype
    of top 10
    structure.
    companies
    from S&P

    Ref10

    Ref11

    Roulette
    Wheel
    Selection

    Based
    on Single point
    comprehensibility crossover
    and
    confidence
    factor
    Based
    on Based
    on Based on
    confidence
    minimum support crossover
    factor
    and
    minimum probability
    confidence factor pc

    Ref12

    Based on
    fitness
    generated

    Based
    on Attribute
    Interesting factor level
    and completeness crossover
    factor

    Ref13

    Individual
    based
    selection
    method

    Based on support
    and confidence
    factor

    Ref14

    Based on
    crossover
    probability
    pc

    Based
    on Individual fitness Random and
    maximal
    based on upgrade heuristic
    crossover
    frequent set index
    thru filtering

    Based on
    mutation
    probability
    pm
    Based on
    mutation
    probability
    pm

    Mutation
    done at
    attribute
    level by
    selecting and
    deleting the
    attribute
    Adaptive
    mutation

    Heuristic
    mutation

    Predictive accuracy

    Number of rules
    based on minimum
    support and
    confidence

    Number of rules
    generated

    500
    Wisconsin
    breast
    cancer
    Balance
    scale
    Chess
    Nursery
    Adult
    Mushroom
    Hayes Roth
    Tic tac toe
    Car
    Evaluation
    dataset

    Mining time

    Student
    achievement
    database
    from schools

    Single table
    generated randomly

    Max
    itemsets
    generated,
    mining time

    Data modification process


    updates the training data by
    checking the mined rules
    and
    By changing the
    representation and they
    grow with generation

    Array representation used


    The crossover and mutation
    operator modified

    User picks maximum


    frequent itemsets. Of
    interest

    operation
    Ref15

    Based
    on
    fitness values
    of individuals

    Based on support
    confidence,
    minimum support
    and confidence

    Random
    selection of
    crossover
    point done
    dynamically

    Genes values
    adjusted
    partially by
    mutation
    Done
    dynamically

    Partial AR after
    certain generation

    Finance
    service data
    from certain
    city

    Ref16

    Random
    selection

    Based on support
    confidence,
    minimum support
    and confidence

    Based on
    crossover
    probability
    pc

    Based on
    mutation
    probability
    pm

    Partial AR after
    certain generation

    Companies
    daily record
    of API

    Distance measure between


    datasets are taken up for
    selection

    Ref17

    Tournament
    Selection

    Based
    on
    comprehensibility
    interestingness
    and
    confidence
    factor

    Randomly
    and based on
    fitness

    No of rules mined

    Adult
    dataset

    Each individual defined as


    expression form

    Ref18

    Based on self
    adaptive
    suppression
    and
    promotion

    If same
    attribute
    present then
    at attribute
    level else
    randomly
    Based on support Grouping
    and
    minimum multipoint
    crossover
    support

    Based on
    mutation
    probability
    pm

    No of rules, reduction Abalone


    of attributes,
    dataset
    execution time

    Rule extracted from final


    population based on
    confidence and minimum
    confidence factor

    You might also like