Professional Documents
Culture Documents
Rahul Sonkamble1, Suyash Patil2, Tejas Dasare3, Saurabh Patil4, Indraneel Pawar5, Aniket Phadtare6
1
Assisstant Professor, Department of CSE, Sanjay Ghodawat Group of Institutions, Atigre, India
2,3,4,5,6
B.E Student, Department of CSE, Sanjay Ghodawat Group of Institutions, Atigre, India
1
sonkamble.rg@sginstitute.in, 2suyashsagar991997@gmail.com, 3itejasdasare@gmail.com,
4
saurabhpatil6468@gmail.com, 5indraneelpawar41@gmail.com, 6aniketphadtare5511@gmail.com
Abstract: In the modern world, sports help to make predictions be more accurate
produce enough considerable statistical and precise than the earlier systems.
information about each player, team, Prediction is the heart of remarkable
games, and seasons. Earlier sports science disciplines in science and that is the reason
was believed to be owned by experts, why philosophy of prediction is employed
coaches, team managers, and analyzers. in many companies. Machine learning
Sports organizations have recently which an area of intelligent systems is will
realized that the availability of science in be used in this report to provide solutions
their data and want to take advantage of to the problems in aspects of predicting
that science through the use of different the results of various sports.
data mining techniques. Sports data
mining assists coaches and managers in
different ways like prediction of the Keywords: Random Forest algorithm
results, performance of the player, performing regression and Database.
identifying the talent, and evaluation of
game strategy. Prediction helps the
managers and the clubs in making the
right decision to win the leagues and the I. Introduction
tournaments. The present study shows
that earlier research on data mining Nowadays, the team managers and
systems to predict the results and evaluate analysts give a prediction about who is
the advantages and the disadvantages of going to win the match. In the present,
each system. Prediction has been different statistics are used for the prediction
successfully applied in all sports. Although of football match result. The prediction of
in many aspects, this application has been any football match is done by using some of
of very small limits. It is very important to the previous data. The prediction of the
look into the applications of the machine
match depends on different variables like
learning in these instances and see if its
player stats, team stats; past data etc. which
application can give better results in the
analysis. By making use of datasets that is is used by club directors and managers to
more precise and machine learning, this decide which team is going to win the match
research aims to offer a solution that will and what will be required to win the match.
In the recent years, football result prediction II. Literature Survey
has gained lots of popularity. Earlier, by
gathering the features that affect the In the whole world football is the most
outcome of football matches, a predictive famous and popular sport. Among all the
model Knowledge discovery in databases sport, football prediction is the most widely
(KDD) was developed. Data mining researched. The main use of our project is to
techniques have also been used in the past. create an application with which we can
In existing systems, statistical and machine predict the upcoming match results based on
learning approaches were used in football many factors. We will give the teams name
match result prediction. For team as the input and it will show which team will
management, the details of the team and the win the match. Previously In the previous
opponent team should be known by the systems they used very less parameters
manager and the other coaching staff which resulted in less accuracy of the
members. The managers and the coaches prediction. We will be using many extra
should be able to identify strengths and factors like player details, half-time score,
weaknesses of different teams and full-time score and many more will be
accordingly prepare for the matches. The considered and the accuracy of the
management should also be able take prediction increases. In the paper Game ON!
decisions during a match and to monitor a Predicting English Premier League Match
match in real time. For all these Outcomes by A. S. Timmaraju [1] they used
requirements, the prediction application a data of just one season and used RBF-
should be able to provide ideas based upon SVM and got accuracy of 60%. The
the current situations. We proposed a model disadvantage of this system is that they used
of football match prediction by using the really less data to predict the outcomes of
few years’ data of every match in English the match. There was an another attempt by
Premier League and trained the algorithm A. Joseph at predicting the match results.
based on that data. The algorithm learns They used Bayesian Nets to predict the
from training data so that it forms different match results but they just got an accuracy
rules and pattern and based on that the ranging 38% and 59%. Finally we were
algorithm makes a decision on some new influenced by Ben Ulmer where they used
data. The algorithm keeps on learning by Random Forest Algorithm and a data of 10
using the feedback with every input. The years to increase the accuracy up to 52% [3].
application predicts the match result. The
results of this application can be used by the
management staff while team selection or by
III. Proposed Methodology
the fans playing while fantasy leagues in
order to create an efficient team or for the
As a quick summary, what we are
purpose of betting.
going to do is take historical results to
calculate the number of goals teams score
and concede. These averages are compared
to the league average and used to create 2. Activity Diagram
values for attacking strength and defensive
strength for every team, which are then At first the user selects the teams and
turned into goal expectation figures. This the match, which is used as input. Then
metric is put into a Poisson Distribution the datasets are checked for the team
formula which works out the probability of details and the data is sent forward for
every result when two teams face each other. Poisson Distribution. After that we get the
We then take these probabilities to create predicted results of the match.
our own odds, compare these against the
bookies’ odds, then identify where there is
value in the market because the bookies are
offering more generous odds that we’d
expect. Simple!
1. Poisson Distribution
(x = Number of goals.
Mean = the probability of that team IV. Implementations and Result
scoring a goal i.e. goal expectancy.
Cumulative = is set to FALSE, so that 1. PhoneGap
the formula returns a value exactly equal
to x. ) [7] PhoneGap is a mobile application
development framework originally
created by Nitobi. PhoneGap enables
software programmers to build
applications for mobile devices using
CSS3, HTML5, and JavaScript instead of everyone is working on the most up-
relying on platform-specific APIs like to-date version.
those in Android, iOS, or Windows
Phone.
Benefits of using PhoneGap: 2. Python
REFERENCES