See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/362538320

ICC T20 Cricket World Cup Prediction Based Data Analytics and Data Mining
Technique

Article · May 2020

CITATIONS READS

0 103

4 authors, including:

Mugdha Rane
Bharati Vidyapeeth's College of Engineering for Women, India, Pune
13 PUBLICATIONS 0 CITATIONS

SEE PROFILE

All content following this page was uploaded by Mugdha Rane on 07 August 2022.

The user has requested enhancement of the downloaded file.

© 2020 JETIR May 2020, Volume 7, Issue 5 www.jetir.org (ISSN-2349-5162)

ICC T20 Cricket World Cup Prediction Based Data

Analytics and Data Mining Technique
Anubha Roy Sakshi Pandey

Priyanka Bhatia PROF. M. A Rane

Department of Information Technology Bharati Vidyapeeth’s College of Engineering for Women,
Katraj, Pune, Maharashtra, India.

Abstract: Keyword:
Melanoma; Data Mining, Machine Learning, Kth
With the advent of statistical modeling in sports,
Nearest neighbor, Naïve Bayes.
predicting the outcome of a game has been
established as a fundamental problem. Cricket is
one of the most popular team games in the world. Introduction:
Statistical modeling has been used in sports for
We embark on predicting the outcome of a One
decades and has contributed significantly to the
Day International (ODI) cricket match using a
success of the field. Cricket is one of the most
supervised learning approach from a team
popular sports in the world, second only to soccer.
composition perspective. Our work suggests that
Various natural factors affecting the game,
the relative team strength between the competing
enormous media coverage, and a huge betting
teams forms a distinctive feature for predicting the
market have given strong incentives to model the
winner. Modeling the team strength boils down to
game from various perspectives. However, the
modeling individual player's batting and bowling
complex rules governing the game, the ability of
performances, forming the basis of our approach.
players and their performances on a given day, and
We use career statistics as well as the recent
various other natural parameters play an integral
performances of a player to model him. Player
role in affecting the outcome of a cricket match.
independent factors have also been considered to
This presents significant challenges in predicting
predict the outcome of a match. We will show that
the accurate results of a game.
the k-Nearest Neighbor (KNN) algorithm yields
The game of cricket is played in three formats -
better results as compared to other classifiers like
Test Matches, ODIs and T20s. We focus our
Naïve Bayes, Support Vector Machine (SVM), etc.
research on ODIs, the most popular format of the
The performance is affected by the type, size and
game. To predict the outcome of ODI cricket
quality of the data.
matches, we will propose an approach where we
first estimate the batting and bowling potentials of
the 22 players playing the match using their career
JETIR2005307 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 35
© 2020 JETIR May 2020, Volume 7, Issue 5 www.jetir.org (ISSN-2349-5162)
statistics and active participation in recent games. a particular feature is independent of the value of
We will use these player potentials to render the any other feature, given the class variable.
relative dominance one team has over the other. Bayesian classifiers are based on Bayes’ theorem.
Taking two other base features into account, Bayes Theorem: Let X be a data tuple and C be a
namely, toss decision and the venue of the match, class label. Let X belongs to class C, then,
along with the relative team strength, we adopt P(C|X) = P(X|C)P(C) / P(X) where;
supervised learning algorithms to predict the • P(C|X) is the posterior probability of class C
winner of the match. The major algorithms used in given predictor X.
the project will be: • P(C) is the prior probability of class.
• P(X|C) is the posterior probability of X given the
Support Vector Machine: class C.
SVM is a supervised machine learning • P(X) is the prior probability of predictor.
algorithm which can be used for both classification
and regression challenges. In this algorithm, we The major contributions of the paper will be:
plot each data item as a point in n-dimensional  We will propose novel methods to model
space (where n is number of features you have) batsmen, bowlers, and teams, using various
with the value of each feature being the value of a career statistics and recent performances of
particular coordinate. Then, we perform the players.
classification by finding the hyper-plane  To predict the winner of ODI cricket
that differentiate the two classes very well. matches, we propose a novel dynamic
Euclidean distance is calculated as the square root approach to react to the changes in player
of the sum of the squared differences between a combinations.
new point (x) and an existing point (xi) across all We will calculate the posterior probability for
input attributes j. each class. The class with the highest posterior
Euclidean Distance(x, xi) = sqrt( sum( (xj – xij)^2 ) probability is the outcome of prediction. For
) this we have to convert the data set into a
frequency table.
Naive Bayes:
Naive Bayes is a simple technique for constructing Related work:
classifiers: models that assign class labels to
Better predictive modeling depends on a better
problem instances, represented as vectors
understanding of the data and attributes selection.
of feature values, where the class labels are drawn
We have to choose between some data mining
from some finite set. There is not a
algorithm. We have chosen data mining as it is
single algorithm for training such classifiers, but a
very flexible in predictive modeling. [5]
family of algorithms based on a common principle:
Prediction, when the game is in progress, is a tough
all Naive Bayes classifiers assume that the value of
task and it needs depending on the best attributes
JETIR2005307 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 36
© 2020 JETIR May 2020, Volume 7, Issue 5 www.jetir.org (ISSN-2349-5162)
that influence the match outcome. Such solutions applications in real sports. Many times news
designed for offline usage and no in-game effects channels organize debates on predictions of cricket
were taken care of. There have been some recent matches. Sponsors and Businessmen invest a lot of
works (20) about in-game decision making to find money on teams without knowing whether their
how much time remaining in the game without team will win or not. Using Machine Learning if
making any prior prediction model. There were we are able to predict the winning team, then it will
several works done in cricket. Bailey and Clarke be easy for the sponsors and other investors to
and Sankaranarayanan [2] used a machine learning think whether they should sponsor the team or not.
approach to predict the result of a one day match It will also become easy to find the places where
depending on the previous data 14 and in-game the team is lagging and this will help the team to
data. Akhtar and Scarf used multinomial logistic work on the particular areas.
regression in their work on predicting an outcome
System Architecture:
of test matches played between two teams.
The project is the developed module that has a
Choudhury [1] used Artificial Neural Network to
User login registration and Admin login
predict the result of a multi-team one-day cricket
Registration. The system finds the Generated
tournament depending on the past 10 years data.
Winning team by using classification when a user-
They used a training set to model the data in a
provided dataset. Admin adds all the information
neural network. Again there were no in-play
related team player as well as a Team also. The
effects that were taken care of. For baseball,
Proposed System is Find out the Predicated
Ganeshapillai and Guttag developed a prediction
Winning Team and generate Result. This system
model that decides when to change the starting
work on two-way working client-server
pitcher as the game progresses. [4] It is very much
similar to our workflow, where they used the
combination of previous data and in-game data to
predict a pitcher's performance. Tulabandhula and
Rudin were designed a real-time prediction and
decision system for professional car racing. The
model decides when are the best time for a tire
change and how many of them. These works
supplied huge encouragement and informative
ideas in our research.

Motivation:

Though Sports Analytics has shown a lot

improvement and advancement but still this
interesting field has been lagging in terms of

JETIR2005307 Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 37

© 2020 JETIR May 2020, Volume 7, Issue 5 www.jetir.org (ISSN-2349-5162)
Generation. players extracted from a particular tournament. We
will observe that simple features can yield very
promising results.

Reference:

[1] A. Aburas, Machine Learning Algorithms for

Big Data Project, Durban: University of Kwa-Zulu
Natal, 2018.

[2] Jesus Maillo, Sergio Ramírez, Isaac Triguero &

Francisco Herrera, kNN-IS: An Iterative Spark-
based design of the k-Nearest Neighbors classifier
for big data, Knowledge-Based Systems 117
(2017) 3–15.

[3] Maryam M Najafabadi, Flavio Villanustre,

Fig. The Proposed System
Taghi M Khoshgoftaar, Naeem Seliya, Randall
1. Registration: The user can register by using Wald, Edin Muharemagic, Deep learning
basic information (First Name, Last Name, Email applications and challenges in big data analytics,
ID, Password Phone Number, etc.) Journal of Big Data volume 2, Article number: 1
2. Login: After Registration, the user can log in. (2015)

3. Classification: Select the team and predict the

winning team by player runs. [4] Weiping Cui, Lei Huang, A map reduce
4. Solution: After classification predicts the solution for knowledge reduction in big data,
winning team. International Journal of Computer Science and
Applications, Vol. 13, No. 1, pp. 17 – 30, 2016

Conclusion:

The paper will conclude the problem of predicting [5] Karl Weiss, Taghi M. Khoshgoftaar, DingDing
the outcome of an ODI cricket match using the Wang, "A survey of transfer learning",Weiss et al.
statistics of 366 matches with the help of KNN J Big Data (2016) 3:9.
classifier and Naive Bayes algorithm. It will devise
method to find the cricket match outcome
prediction, team structure analysis and player
recommendation system using the statistics of the

JETIR2005307
View publication stats
Journal of Emerging Technologies and Innovative Research (JETIR) www.jetir.org 38