Docstoc

A MODEL FOR TALENT IDENTIFICATION IN CRICKET BASED ON OWA OPERATOR

Document Sample
A MODEL FOR TALENT IDENTIFICATION IN CRICKET BASED ON OWA OPERATOR Powered By Docstoc
					   INTERNATIONAL JOURNAL OF & Management Information System (IJITMIS),
 International Journal of Information TechnologyINFORMATION TECHNOLOGY &
 ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME
               MANAGEMENT INFORMATION SYSTEM (IJITMIS)

ISSN 0976 – 6405(Print)
ISSN 0976 – 6413(Online)                                                     IJITMIS
Volume 4, Issue 2, May - August (2013), pp. 40-55
© IAEME: www.iaeme.com/ijitmis.html
Journal Impact Factor (2013): 5.2372 (Calculated by GISI)                 ©IAEME
www.jifactor.com




  A MODEL FOR TALENT IDENTIFICATION IN CRICKET BASED ON
                     OWA OPERATOR

                     1
                         Gulfam Ahamad, 2S. Kazim Naqvi , 3M.M. Sufyan Beg
  1
      FTK-Centre for Information Technology, Jamia Millia Islamia, New Delhi – 110025, India.
  2
      FTK-Centre for Information Technology, Jamia Millia Islamia, New Delhi – 110025, India.
      3
        Department of Computer Engineering, Jamia Millia Islamia, New Delhi – 110025, India.


 ABSTRACT

         Talent identification in sports is a challenging and significant task which is considered
 highly subjective. However, several attempts have been made in the past to reduce the
 subjectivity in this task. In this paper we have reviewed several talent models which have been
 proposed in the past. We have also presented a brief summary of each of these models
 focusing on their modus-operandi. The paper also identifies essential parameters for talent
 assessment in cricket. A model based on Ordered Weighted Averaging Aggregation (OWA)
 operator has also been proposed. The paper also presents an example demonstrating the
 application of the proposed algorithm on sample data.

 Keywords: Talent identification, OWA, Linguistic Quantifier, Talent Classifier.

 1.0      INTRODUCTION

         The ability to perform well in sports may vary amongst individuals. A person may be
 exceptionally good at one sport whiles he may only be average at others. It is interesting to
 study what determines the ability of a person to excel or not excel in a given sport. This ability
 is commonly referred to as talent. Talent refers to the skills that someone has naturally to do
 something that is hard. A number of authors [20], [19], [14] have defined talent as an
 increasable natural endowment of a superior quality of a person.
         Talent identification is a process to identify the ability of superior quality. It is a
 complex multifaceted, multidimensional and multi-stage process [5] [11] [17] [28]. Many
 earlier studies [19] [12] [29] have characterized the talent by a number of factors viz. Health,
 Motor, Functional, Morphological, Physiological, Anthropometric, Psychological, Social,
 Cultural, Game Intelligence, Technical/ Tactical Abilities and Genetics. Although, no



                                                40
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

consensus seems to have emerged on the completeness of the above parameters which are
believed to contribute towards talent in an individual.
        Talent identification is another area with significant importance for individuals and
sports organizations. Correct and timely identification of sports talent can build careers and
bring glory to the nations. On the other side, persistence on incorrectly chosen sports will
invariably lead to wastage of time and resources. To identify talent in sports a number of
studies have been made in the past [8] [18] [29] [1] [21] [28] [9] [6] [6] [5] [26] [27] [26] [3]
[2] [13]. During Our literature survey, we did not come across any study for Talent
Identification in Cricket.
        In this paper, we have proposed a model based on OWA Operator to assess the talent
of a cricket enthusiast. In section 1.0 of the paper, we summarize the frequently cited studies
in Talent Identification. The studies summarized in section 1.1 also conclude that no Talent
Identification Model has been developed for Cricket. In section 2.0, we present the parameters
which can be effectively used for building talent identification model for cricket. In section
3.0, we discuss OWA operator. Applications of OWA operator for talent identification in
cricket has been discussed in section 4.0. In Section 5.0, we present the conclusions.

1.1     Talent Identification Models in Sports
        Timely talent identification in sports is very significant and challenging task. To
identify talent in sports, various computer based talent identification models have been
proposed. These models could potentially play a very significant role in lives of sports
enthusiast and sports organizations. Some of these are listed below [7]:

      • An individual can make timely decision to pursue a sport of his interest or not.
      • Provide a cost effective way for assessing talent levels without assistance of coaches.
      • They can help in increasing the confidence levels in athletes.
      • Online talent models can even be more potent as they provide unparalleled reach. They
        can be accessed anytime anywhere by anyone.
    • Online Talent Identification models can also help sports authorities in getting
        indications on talent distribution in various geographical regions within the country.
        Such model can help in separating extraordinarily talented athletes from average.
    • They can also address the challenge of inadequate number of coaches available in the
        country.
        All Talent Identification models leverage physical characteristics of athletes which
they deem significant for the sport. Most of the models have applied statistical techniques
including variance [8], standard deviation, t-test, regression [3] [27] [6] [29], ANOVA [9],
MANOVA [26] [4] [21], Few models have also attempted to use Fuzzy Logic and Expert
Systems. Also majority of the models have been developed for Track & Field sports [1] [18]
[8], a few of them have been developed for soccer [21] [24] [28], hockey [4], water polo [6],
baseball [3] [2], handball [26] and table-tennis [13]. Here we present a brief summary of 15 of
such models.
        In [8], a model was proposed to enable identification of talent in track and field sports
in Iran. The model chose the age range of 6-12 years. The basic parameters employed by the
model are motor ability, anthropometric, psychological, physiological, sociological and
cultural characteristics of prospective athlete. The model used statistical variance on the
identified parameters.



                                               41
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

         A model for 14 identified sports (athletics sprint jump, martial arts of kicking type,
martial arts of pulling and pushing type, football, tennis, handball, volleyball, water polo,
rowing, swimming, athletics long distance running, basketball, athletic throwing and
gymnastics) was proposed [18]. The model was applied on the athletes in the age range of 6-
18 years. The model uses characteristics which include: motor skills, morphological and
functional and it uses expert system and fuzzy logic membership technique to help rate talent
of a person in a particular sport.
         In [29], Model was presented an application for selection of sports talent using
statistical regression equation with computer programming is proposed for athletic jump
without any age criteria. Authors have claimed that their approach can also be applied to other
sports as well.
         In [1], a model was proposed for the age range of 11-15 years with the characteristics
of sports interactive task, physiological capacities, motor capacities and biometrics qualities. It
was implemented on Scottish children with the help of statistical techniques. The model was
meant to determine talent of children in 12-identified sports (high jump, long jump, sprinting,
hurdling, karate, triathlon, shot-put, skiing (DH), curling, hockey, tennis and squash)
         In [21], a model was proposed for the age range of 15-16 years with anthropometric,
physiological, psychological and soccer specific characteristics. It used Multivariate Analysis
of Variance (MANOVA) technique to distinguish between elite and sub-elite groups on the
basis of performance on test items.
         In [28], a model was proposed to determine the relationship between physical and
performance characteristics for the age of 13-16 years. The authors applied MANCOVA
technique to the data to identify the most important physical parameters for soccer.
         In [9], a model predicted talent of players in age range of 18-years. The model used
ANOVA statistical methods on the physical characteristics of player which include body
height, weight, skeletal age, choice reaction time, stepping speed and stepping endurance.
         A study [6] was meant to identify and develop talent in water polo sports. It was
proposed for the age range of 14-15 years with the characteristics of motor ability related to
water activity, physical ability and evaluation of game intelligence. The authors applied
statistical analysis unpaired t-test ANOVA on the data to assess the talent.
         This research [5] demonstrated application of MANCOVA to understand the
relationship between multidimensional performance characteristics and level of performance
in talented youth hockey players. The model was implemented for the age range of 13.2-14.2
years and considered the anthropometric, physiological, psychological, technical and tactical
characteristics.
         In the [16], a model was meant to detect talent in handball sports. The model used
MANCOVA on morphological, physical fitness, anthropometric, hand ball specific motor
skills and maturation characteristics. The model was applied on group of 14-16 years.
In [27] a model used essential characteristics of team sport. The model applied regression and
canonical analysis techniques.
         A method in [28] is proposed for selecting players in team sports game with the help of
standard deviation and expert system.
         In [3], a model was proposed for Croatia. The model uses characteristics which include
potential, morphological, technical knowledge and coordination abilities with the help of
statistical techniques t-test, z-test, and correlation.
         This was proposed for different age groups with the help of graphical statistical
techniques. The model used characteristics which included special tactical and technical skills
for basketball [2].

                                               42
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

         In [13], a model was developed for children in the age range of 6-8 years with the help
of t-test with standard deviation. The model uses characteristics including height, weight, skin
fold, deep bend for flexibility, polygon for coordination, bent arm hang on horizontal bar for
strength test, sit up test for trunk strength, standing jump for explosive strength, 60 meters
sprint for speed and 600 meters run for endurance.
Based on our survey of published literature we can infer that no model for identifying of talent
in cricket has been reported as on date. Further, to develop TID model for cricket we need to
identify parameters which can be measured and thus can produce data for analysis. In the next
section we identify such parameters which may play important role in identifying talent in
Cricket.

2.0    PARAMETERS FOR TALENT IDENTIFICATION IN CRICKET

       The Talent Identification report [25] summarizes the talent parameters for cricket. The
parameters are based on physical/motor, anthropometric and cognitive characteristics. To
quantify these parameters various tests have been identified in [7].The identified
characteristics and its parameters are listed in Table1 and Parameters with tests are listed in
Table2.


                Talent Characteristics                Talent Parameter


               Physical/ Motor Ability      Speed, Agility, Flexibility, Balance
               Tests                        Static/ Dynamic, Endurance, Upper
                                            Body Strength, Lower Body
                                            Strength, Fatigue Index, Shoulder
                                            Flexibility, Bowler Accuracy, Under
                                            Arm Throw Accuracy, Under Arm
                                            Throw Accuracy, Catching Ability,
                                            Ground Fielding Ability

               Cognitive Ability Tests      Self Motivation, Reaction Time,
                                            Hand Eye Coordination, Creativity,
                                            Decision Making, Self Control &
                                            Self Monitoring, Integrity and Work
                                            Ethic, Willingness, Concentration
                                            and Focus, Stress

               Anthropometric Tests         Body Mass Index, Vo2max

 Table 1: Talent requirements for cricket in terms of various parameters with characteristics




                                               43
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME


           Parameters                                        Tests
  Speed                           Speed Test (T1)
  Agility                         Illions Agility Test (T2)
  Endurance                        Step Up and Down Test (T3)
  Stress                          Stress Test Quiz (T4)
  Self Motivation                 Self Motivation Test Quiz (T5)
  Upper Body Strength             Push Up Test (T6)
  Lower Body Power                Hop Run Test (T7)
  Reaction                         Ruler Catching Test (T8)
  Flexibility                     Sit and Reach Test (T9)
  Fatigue Index                   RAST (Running Based Anaerobic Sprit) Test (T10)
  Bowler Accuracy                 Bowler Accuracy Test (T11)
  Through Catching Accuracy       Through Catching Accuracy Test (T12)
  Under Arm Through                Under Arm Through Accuracy Test (T13)
  Accuracy
  Catching Ability                Catching Ability Test (T14)
  Ground Fielding                 Assessment of Clean Pick Ups. (T15)
  VO2 Max                         Maximum Oxygen Up Taken Test (T16)
  Body Mass Index                 Weight/ Height2 (T17)
  Hand Eye Coordination           Catching and Throwing the Ball in Cyclic Order with
                                  Hands (T18)
  Creativity                      Creativity Test Quiz (T19)
  Decision Making                 Decision Making Ability Test Quiz (T20)
  Self Control and Self           Self Control and Self Monitoring Test Quiz (T21)
  Monitoring
  Will Power                      Will Power Test Quiz (T22)
  Self Confidence                 Self Confidence Test Quiz (T23)
  Integrity and Work Ethic        Integrity and Work Ethic Ability Test Quiz (T24)
  Shoulder Flexibility            Shoulder Flexibility Test Based On Physical Action (T25)
  Balance                         Beam Test for Balance (T26)
  Balance in Static Form          Balance Test Based on Physical Action (T27)
  Concentration and Focus         Concentration and Focus Skill Test Quiz (T28)

      Table 2: Talent requirements for cricket in terms of various tests and parameters.


       Various quantitative tests (given in table1) have been identified in [7] which can help
in measuring each one of the above characteristics. Given the ability to quantify the
characteristics for talent assessment in Cricket, a talent identification model can be build. In
section 4.0 we first introduce Ordered Weighted Averaging/Aggregation operator, which we
apply to the problem of Talent Identification in section 4.0.



                                               44
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

3.0      ORDERED WEIGHTED AVERAGING AGGREGATION (OWA) OPERATOR

        The OWA operator was introduced by [22] to provide a means of aggregation, which
unifies in one operator the conjunctive and disjunctive behavior. It provides a parameterized
family of aggregation operators including many of the well-known operators like maximum,
minimum, k-order statistics, median and arithmetic mean. For n different scores x1 , x2 ,..., xn ,
the aggregation of these scores may be done using the OWA operator as follows.

                                              n
                OWA ( x1 , x2 ,..., xn ) =   ∑w y
                                             i =1
                                                    i   i




where y i is the ith largest score from amongst x1 , x2 ..., xn . The weights are all non-negative
                 n        
∀i, wi ≥0, and  ∑ wi = 1 . We note that the arithmetic mean function may be obtained using
                 i=1      
                                    1
the OWA operator, if i, wi = . Similarly, the OWA operator would yield the maximum
                           ∀        n
function with wi =1 and wi =0 for all i≠ 1. The minimum function may be obtained from the
OWA operator when wn =1 and wi = 0 for all i ≠ n.
In fact, it has been shown [22] that the aggregation done by the OWA operator is always
between the maximum and minimum. To find the values of the weights wi , we need to make
use of the relative fuzzy linguistic quantifiers, explained as follows.

3.1      Relative Fuzzy Linguistic Quantifier

A relative quantifier, Q: [0, 1] → [0, 1], satisfies:
             Q(0) = 0,
             ∃r ∈ [0,1] such that Q(r) = 1.
In addition, it is non-decreasing if it has the following property:
         ∀a, b ∈ [0,1] , if a> b, then Q(a) ≥ Q(b).
The membership function of a relative quantifier can be represented as shown in [1]:
            0
            r − a           if r < a
            
    Q(r ) =                 if a ≤ r ≤ b , …………..[1]
             b−a
                             if r > b
            1
            
where a , b, r ∈ [0,1] and Q(r ) = Q(i / m)

[22] Computes the weights wi of the OWA aggregation from the function Q describing the
quantifier. In the case of relative quantifier, with m criteria [23],

      wi = Q (i / m) − Q ((i − 1) / m)), i = 1,2,...., m, with Q(0) = 0.




                                                            45
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

3.2     Normalized Adequacy Coefficient
         The adequacy coefficient [15] is an index used for calculating the differences between
two elements, or two sets, or two fuzzy sets, etc. The adequacy coefficient is very similar to
the hamming’s distance but with some differences. It makes it more complete in a lot of
decision making problems, especially, when we cannot accept that one set (Y) is higher than
the other (X). The similarity between two sets in the adequacy coefficient is calculated with
(1 ∧ (1 − x + y ) ) or the complement (0 ∨ ( x − y ) ) . For two sets X = {x1 ,..., xn } and
Y = { y1 ,..., y n } , the weighted adequacy coefficient can be defined as follows.

Definition: According to [15], A weighted adequacy coefficient of dimension-n is a mapping
WAC: [0,1] × [0,1] → [0,1] that has an associated weighting vector W of dimension n with
           n      n

 n

∑w
j =1
       j   = 1 and w j ∈ [0,1] , such that:
                                          n
        WAC ( x1 , y1 ,..., xn, yn ) = ∑ wi [1 ∧ (1 − xi + yi )] ,
                                         i =1

where xi and yi are the ith arguments of the sets X and Y, respectively.
Note that if wi = 1 / n , for all i, then, the weighted adequacy coefficient becomes the
normalized adequacy coefficient (NAC).
                                 1 n
   NAC ( x1 , y1 ,..., xn, yn ) = ∑ [1 ∧ (1 − xi + yi )] ,
                                 n i =1

4.0  APPLICATION OF OWA TO TALENT IDENTIFICATION PROBLEM IN
CRICKET

        As discussed in section 2, the talent identification is a difficult task. Attempts have
been made to identify parameters to assess talent in Cricket. These parameters were
summarized in section 3. To build a cricket talent assessment model, we collected responses
from cricket experts for the identified 28-tests. The experts were requested to categorize the
result range of each of the 28-tests to judge the suitability of the result value for each of the
five talent classes, viz. Extraordinary Talented [EOT], Very Much Talented [VMT], Much
Talented [MT], Moderately Talented [MDT], and Not Talented [NT]. A sample question to
record assessment of an expert is shown in figure1:

Please fill your rating for the "Speed Test" for the five categories of talent classes.
Run 35 meters in a straight line and record the time in seconds.

                              [<4.80 sec ] [4.80- 5.09 sec] [5.10-5.29sec] [5.30- 5.60 sec ] [ >5.60 sec ]
 Extraordinary Talented
 Very Much Talented
 Much Talented
     Moderately Talented
 Not Talented
                             Figure1: A sample question from questionnaire.

                                                      46
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

        It may be observed that expert is expected to record his assessment for each of the 5-
talent classes by choosing an interval. For each of the chosen interval we record its average
value as the assessment vet[i] of the expert. All such assessments for each talent class-i are
recorded in an Expert Assessment EA[i] matrix as follows:


                   t1       t2        . . . t 28
         e1  ve1t1 [i ] ve1t2 [i ]   . . . ve1tm [i ]
                                     . . . ve2tm [i ]
         e2 ve2t1 [i ] ve2t2 [i ]                    
          .  .              .        . . .     . 
 EA[i] =                                              ………. (2)
          .  .              .        . . .     . 
          .  .              .        . . .     . 
                                                     
         en vent1 [i ] vent2 [i ]
                                     . . . ventm [i ]
                                                      


∀j,1 ≤ j ≤ 28. t j is a test aimed at quantifying the physical ability.

where i ∈ [0,1,2,3,4 ]
i = 0 for EOT
i = 1 for VMT
i = 2 for MT
i = 3 for MDT
i = 4 for NT

Test Definition: Each test t i is a 5x5 matrix comprising of rows representing Talent Classes
(TC) and columns representing permissible responses. Each response is recorded in form of
(a, b) where a is the lowest permissible value and b is the highest permissible value for each
class and each test. Thus,

    t i = TCR (TC , Re sponse)∀i ∈ [1,28].

    Where, TCR means Talent Class Response Matrix.

The data for all the values obtained in above 5-matrices [eqn. 2] were normalized using min-
max transformation so that each value now falls in [0, 1]. Min-max normalization maps a
value v of an attribute A to v′ in the range of [new_minA, new_maxA] by computing.

             v − min A
    v' =                 (new _ max A − new _ min A) + new _ min A .
           max A − min A

The five normalized matrices EA[i] now comprise the Knowledge Base for our Model. A
broad architecture for the Talent Identification Model is shown in figure 2. Any cricket

                                                    47
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

enthusiast can provide the outcome result values for himself for each of the 28-tests. The
model then uses the Talent Classifier described in section 4.2 to classify the talent of person in
any one of the five possible talent classes.



                                       Cricket Enthusiast


                       o1   o2 . . .     . . oi . .       .   .   .   . o28




                             Normative Data for Talent (NDFT)
                                                                                  EOT

                                                                                  VMT
                                            Talent
                                        Classifier (TC)                           MT

                                                                                  MDT
                                       Knowledge Base
                                                                                  NT
.

                  Figure 2: Architecture of Cricket Talent Identification Model

4.1    Algorithm for Talent Identification
The algorithm for talent identification is given below:

Algorithm to compute weights (wi)
begin
for each talent class TC[ j ]; j ∈ [0,4] .
    for each test t i , i ∈ [1,28]
        for each expert E (k ); k ∈ [0, n] where n = total no. of experts
             r (k ) = k / n;
        for each expert E (k ); k ∈ [0, n]
            begin
            if r (k ) < a
                         QR(k ) = 0
               else if {a ≤ r (k ) ≤ b}
                             QR(k ) = (r (k ) − a) /(b − a)
                   else
                           r (k ) > b}


                                                 48
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

                          QR(k ) = 1
              end;
                  if (k <> 0)
               w[k , i] = QR(k ) − QR(k − 1);
      End Loop k;
   End Loop j;
End Loop i;

//For the Talent Class TC[ j ]; j ∈ [0,4] and test t i , i ∈ [1,28] calculate OWA [j,i] as follows:
    for each talent class TC[ j ]; j ∈ [0,4] .
        for each test t i , i ∈ [1,28]
            for each expert E (k ); k ∈ [1, n] where n = total no. of experts
            if a < b then
                 y(k) = arrange_and_assign(Π ti ( EA[ j ]), ascending);
            else
                 y(k) = arrange_and_assign(Π ti ( EA[ j ]), descending);
        end for k;
        sum: = 0;
        for each expert E (k ); k ∈ [1, n] where n = total no. of experts

          sum += w(k,i)*y(k);
       OWA(j, i) = sum;
   End for j;
End for i;
End;

4.2    Classification of Talent
        To classify the talent of a person, we first record his/ her test records, oi for each of the
28-tests in a 28x1 NDTF (Normative Data for Talent) matrix as follows:
NDFT (v) = (o1 , o2 ............o28 )
We now calculate the normalized adequacy coefficient (NAC) between NDFT and OWA(j,i)
where j ∈ 0..4 and i ∈ 1..28 . The classifier now classifies the talent based on the
Max( NAC( NDFT , OWA( j, i ))), j ∈ 0..4, i ∈ 1..28 . For example, if the maximum value is
obtained for j=3, then Talent class is identified as “Much Talented”.

4.3    Experiments and Results
        We now demonstrate the application of algorithm on the ith test viz. the speed test and
     th
the j talent class, viz. “Extra Ordinary Talent (EOT)”. Firstly, we normalize the rough range
for the identified test for the five talent classes. The sample normalized values are shown in
figure 3. We assume that assessments of 33 experts (E1, E2, E3,….., E33) are available and
recorded in the matrix Experts Assessment (EA) as described in equation 2.




                                                  49
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME




     1




         0                                                                            1   x
             0.0-0.465   0.46-0.515      0.52-0.549         0.55-0.60     0.601-1.0
              EOT          VMT              MT              MDT              NT

                          Figure3: Relative Fuzzy Linguistic Quantifiers


        For the jth talent class, with the sample normalized interval a= 0, b= 0.465 we now
calculate weights using algorithm described in section 4.1. This is also depicted in table 3
below:


 i= 0,1,2,…..33                         0             1           2            …              33
 i/33 = r                             0/33= 0    1/33          2/33            …          33/33
                                                 =0.03         =0.06                       =1

 Q(i / n) = Q(r )
         0
          r − a if r < a
                                       0            0.06       0.13           …              1
 Q(r ) =         if a ≤ r ≤ b
          b−a
                  if r > b
         1
         




        i       i −1                 Not         0.06-0       0.13-0.06        …           1-1
 wi = Q( ) − Q(      )             Required                                               =0.000
        n         n                              = 0.06         =0.07


                                 Table3: Calculation of Weights




                                                50
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME




                      Table4: Weights for Extra Ordinary talented (EOT)




                  Table5: Experts opinion for Extra Ordinary talented (EOT)

                                               51
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

       Now the ordered weighted averaging aggregation (OWA) operator is used to aggregate
the experts opinions for EOT class.
                 33
OWA(0,1) =      ∑i =1
                         [0.06, 0.07, 0.06,……….. 0.00 ]. Ascending or Descending [15] [0.23, 0.22,

0.33, ….… , 0.23) = 0.89.

        Similarly, values are aggregated for all the tests. These values are depicted in Table 6.
The normative data for talent (NDFT) of cricket enthusiast for all tests is depicted in Table7.
Now we find the Normalized Adequacy Coefficients using the equation (3) for NDFT of
cricket enthusiast with all five talent class.




                         Table6: Aggregated values of all talent classes for 28 tests




                        Table7: Assessed results of cricket enthusiast against 28 tests

Normalized Adequacy Coefficient value for Extraordinary Talent (EOT):
                                1 n
NAC ( x1 , y1 ,..., x n, y n ) = ∑ [1 ∧ (1 − xi + y i )]..............(3)
                                n i =1
NAC (VMT − Enthusiast) = NAC ( 0.219,0.489 , 0.23,0.54 , 0.041,0.28 ,......, 0.9,0.85 )
    1
=      [1 ∧ (1 − 0.219 + 0.489) + 1 ∧ (1 − 0.23 + 0.54) + ............ + 1 ∧ (1 − 0.9 + 0.85)].
    28
 1
    [1 ∧ 1.27 + 1 ∧ 1.31 + ......... + 1 ∧ 0.95]
28
   1
= [1 + 1 + ............ + 0.95]
   28
   1
= [ 24.81] = 0.8860714
   28
Similarly, the values for other classes are calculated and shown below.
NAC(VMT-Enthusiast)= 0.9209286
NAC(MT-Enthusiast)= 0.8971786
NAC(MDT-Enthusiast)= 0.8983571
NAC(NT-Enthusiast)= 0.8206071

                                                      52
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

Now, Max Operator is used to find the maximum value among the above given normalized
adequacy coefficients.
MAX (NAC(EOT-Enthusiast), NAC(VMT-Enthusiast), NAC(MT-Enthusiast), NAC(MDT-
Enthusiast), NAC(NT-Enthusiast))
= MAX(0.8860714, 0.9209286,0.8971786, 0.8983571, 0.8206071) = 0.9209286
Since the maximum Max(NAC(TC[j], NDFT)) ∀ j ∈ (0,4) is achieved against VMT. We
identify the talent of enthusiast as “Very Much talented”.

5.0    CONCLUSION

        This paper reviewed several talent identification models in sports which have been
proposed in research papers. The study revealed that no talent identification model has yet
been proposed for cricket. To develop such a model we identified 28-parameters and
corresponding tests to quantify the talent of an enthusiast against these parameters. We then
build a database of experts’ opinion for classification of talent into five categories based on the
results of the identified tests. An algorithm using Ordered Weighted Averaging Aggregation
(OWA) operator was proposed for aggregation of opinions whose results can be easily used
for classification of talent in cricket. The paper also presents an experiment with results to
demonstrate the application of the model.

REFERENCES

[1]. Abbott A. and Collins D., “A Theoretical and Empirical Analysis of a ‘State of the Art’
     Talent Identification Model”, Taylor & Francis Group of Publication, Vol 13, Number
     2, 1 Dec- 2002, pp. 157-178(22).
[2]. Dezman B., Slavko Trninic and Drazan Dizdar “Expert model of Decision making
     System for efficient Orientation of Basketball players to positions and roles in the
     game- Empirical Verification” Coll. Antropol, 25(2001) 1: 141-152.
[3]. Dezman B., Slavko Trninic and Drazan Dizdar, “Models of expert system and decision-
     making systems for efficient assessment of potential and actual quality of basketball
     players”, Kinesiology (Zagreb), 33, 2:207-215.
[4]. Doug MacCurdy, “Talent Identification around the World and Recommendations for
     the Chinese Tennis Association” International Tennis Federation Coaching.
     AvailableOn:http://www.itftennis.com/shared/medialibrary/pdf/original/IO_18455_orig
     inal.PDF
[5]. Elferink-Gemser MT, Chris Visscher, Koen Lemmink et.al. “Relation between
     multidimensional performance characteristics and level of performance in talented
     youth field hockey players”, Journal of Sports Sciences, 2004 Nov-Dec; 22(11-12):
     1053-63.
[6]. Falk B., Ronnie Lidor, Yael Lemdr et al. “Talent Identification and early Development
     of Elite Water-polo players: a 2 year follow-up study”, Journal of Sports Sciences, 2004
     Apr; 22(4): 347-55.
[7]. Gulfam Ahamad, S.K.Naqvi and M.M. Sufyan Beg, “A study of talent identification in
     models in sports and parameters for talent identification in cricket”, International
     Conference on Physical Education and Sports Sciences, November 16-18, 2012.



                                               53
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

[8]. Hadavi F. and Zarifi A., “Talent Identification and Development Model in Iranian
      Athletics” World Journal of Sport Sciences, Iran, 2009, vol. 2(4), pp. 248-253, ISSN
      2078-4724.
[9]. Hirose Norikazu, “Prediction of talent in youth soccer players: prospective study over
      4-6 years”, Football Science, Vol. 8, 1-7, March, 2011.
[10]. http://simple.wikipedia.org/wiki/Talent
[11]. Kate Vrljic & Mallett Clifford J., “Coaching knowledge in identifying football talent”
      International Journal of Coaching Science, 2008, Australia.
[12]. Kluka Darlene A., “Long-Term Athlete Development: Systematic Talent
      Identification”. Available on http://www.bso.or.at/fileadmin/Inhalte/Dokumente/Internationales/EU
      Study Training Addendum.pdf.
[13]. Kondric Miran “The expert system for orientation of children in to table tennis in the
      Republic of Slovenia”, International Journal of Table Tennis Sciences, No. 3(1996).
[14]. Mallillin, Tiffany R. and Josephine Joy Reyes B., et al, “Sports Talent Identification in
      1st and 2nd year UST High School Students” Philippine Journal of Allied Health
      Sciences, November 2007, vol. 2, Issue 1, pp. 41-42.
[15]. Merigo’, J.M. & Gil-Lafuente, A.M. (2011). Decision Making in Sport Management
      Based on the OWA Operator. Expert System with Applications, 38, 10408-10413.
[16]. Hasan Mohamed, Rael Vaeyens and Stijin Matthys et al, “Anthropometric and
      performance measures for the development of a talent detection and identification
      model in youth handball”, Journal of Sports Sciences, Vol. 27, Issue 3, 4 Feb 2009, pp.
      257-266.
[17]. Oleksandr Karasilshchikov, “Talent Recognition and Development- Elaborating on a
      Principle Model” International Journal of Developmental Sport Management, Kubang
      Kerian, 16150, Kelantan, Malaysia, Vol. 1(1) 2011 1.
[18]. Papic Vladan, Nenand Rogulj and Vladimir Plestina, “Identification of sport talents
      using a web oriented expert system with a fuzzy module.” Elsevier Journal, Expert
      System with Applications, Split (Croatia), 2009, vol. 36, pp. 8830-8838.
[19]. Pearson D.T., G.A. Naughton and M. Torode, “Predictability of physiological testing
      and the role of maturation in talent identification for adolescent team sports.” Journal of
      Science and Medical in Sport, 17 July 2006, Vol. 9, issue: 4, pp. 277-287.
[20]. Peltola, 1992; Williams & Reilly, 2000, “Talent Identification in British Judo, pp. 216,
      AvailableOn:www.bath.ac.uk/sports/judoresearch/Full%20texts/Talent%20Identificatio
      n%20in%20British%20%20Judo.pdf
[21]. Reilly T., Williams A. M., Nevill A. & Franks A. (2000): A multidisciplinary approach
      to talent Identification in soccer, Journal of Sports Sciences, 18:9, 695-702.
[22]. R.R. Yager, On ordered weighted averaging aggregation operators in multimedia
      decision making, IEEE Trans. Systems Man Cyber net. 18 (1988) 183-190.
[23]. Sheng-Lin Chang, Reay-Chen Wang and Shih-Yuan Wang, “Applying fuzzy linguistic
      quantifier to select supply chain partners at different phases of product life cycle”,
      International Journal of Production Economics, Vol. 100, Issue. 2, April 2006, pp. 348-
      359.
[24]. Spinks W., “e-book on Science and Football -iv”, Routledge, Taylor and Francis Group.
[25]. Talent       Identification      Report      Explanatory       Notes,     available     on
      “static.ecb.co.uk/files/talent-id-report-explanatory-notes-1403.doc”
[26]. Slavko Trninic, Vladan Papic and Victorija Trninic, “Player Selection Procedures in
      Team Sports Games”, Acta Kinesiologica (2, May, 2008) 1: 24-28.


                                                 54
International Journal of Information Technology & Management Information System (IJITMIS),
ISSN 0976 – 6405(Print), ISSN 0976 – 6413(Online) Volume 4, Issue 2, May - August (2013), © IAEME

[27]. Slavko Trninic, Igor Jelasa and Vladan Papic, “Global nonlinear model for efficacy
      evaluation in team Sports”, Journal of Sports Sciences, 2(2009) 2: 73-80.
[28]. R .Vaeyens, R.M. Malina and M Janssens, “A Multidisciplinary Selection Model for
      Youth Soccer: The Ghent Youth Soccer Project” British Journal of Sports Medicine,
      2006; 40: 928-934.
[29]. Yu Xin, Zhiqiang Ye and Lingnan Mei, “Application of computer in selection of sports
      talents based on regression equations.” Knowledge Acquisition and Modeling
      Workshop, 2008. IEEE International Symposium on 21-22 Dec. 2008, Wuhan, pp. 697-
      699, Print ISBN: 978-1-4244-3530-2.
[30] Lokesh S. Khedekar and Dr.A.S.Alvi, “Advanced Smart Credential Cum Unique
      Identification and Recognition System(ASCUIRS)”, International Journal of Computer
      Engineering & Technology (IJCET), Volume 4, Issue 1, 2013, pp. 97 - 104, ISSN Print:
      0976 – 6367, ISSN Online: 0976 – 6375




                                               55

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:0
posted:7/6/2013
language:
pages:16