e-ISSN: 2582-5208

International Research Journal of Modernization in Engineering Technology and Science

( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:10/October-2023 Impact Factor- 7.868 www.irjmets.com

USING MACHINE LEARNING TO PREDICT THE WINNER OF

THE CRICKET MATCH
Sharad Thavale*1
*1UG Student, Department of Information Technology, B.K Birla College Kalyan,
(Empowered Autonomous Status), Kalyan, Maharashtra, India.
ABSTRACT
In this paper, we predict the performance of the cricket team will win or lose depends not only on the
performance of the team but also on the pitch (bowling condition), batman strength, and last score on this
ground, so datasets have been made use for the train model Naïve Bayes, Decision Tree Regression, Support
Vector Classifier (SVC), Random Forest Classifier and train all the two Datasets. The finding prompted a
revolution of the balance between traditional prediction and advanced techniques to find the output of winning
team chances. The evaluators the performance of each team’s players by calculating accuracy and generating
confusion matrices to visualize classification results. Each model has a different accuracy.
I. INTRODUCTION
Cricket, second of the most popular sports in the world, has a massive fan following and garners attention from
enthusiasts, bookmakers, and analysts alike. In machine learning technology gathering cricket data and
building a prediction model for a better understanding of the cricket game. The condition of the pitch can favor
either batsmen or bowlers and weather conditions can affect the game like swing for fast bowlers or turn for
spinners. This machine learning model is trained with the help of a dataset of the stadium, the team’s last
performance, for giving good accuracy prediction. Predicting cricket match outcomes involves a combination of
historical data analysis, team and player assessment, pitch and weather conditions, and various other factors.
It's a complex process that combines both quantitative and qualitative aspects, and no prediction is ever
foolproof due to the inherent.
Cricket, often referred to as a religion in some parts of the world, captivates millions of fans with its thrilling
contests and unpredictable outcomes. The sport’s intricacies, including player performance, pitch condition,
and team dynamics, make predicting cricket match winners a challenging yet intriguing task. In recent years,
the fusion of sports analytics and machine learning has provided a promising avenue for enhancing our ability
to forecast the outcome of cricket matches. Cricket’s popularity extends far and wide, with passionate fans
following the sport across the globe. For cricket enthusiasts, the excitement of watching matches is often
heightened when they have a stake in predicting the winner. This extends to sports analysts, bookmakers, and
even cricket teams who seek data-driven insights to strategize effectively. Cricket is a complex sport influenced
by numerous factors. The outcome of a match depends on the form of individual players, team dynamics, pitch
conditions, and even the impact of weather. These intricate variables pose a formidable challenge when
attempting to predict match results using conventional methods. Machine learning, a subset of artificial
intelligence, offers a data-driven approach to overcoming the challenges of predicting cricket winners. By
analyzing historical data, identifying patterns, and learning from vast datasets, machine learning algorithms
have the potential to provide valuable insights and accurate predictions. Machine learning (ML) is a
transformative field of artificial intelligence that empowers computers to learn from data and make predictions
or decisions without being explicitly programmed. This abstract explores the essence of machine learning, its
application, and its profound impact on various domains. At its core, machine learning leverages algorithms
and statistical models to identify patterns and relationships within vast datasets. These algorithms and
statistical models identify patterns and relationships within vast datasets. These algorithms enable machines to
adapt and improve their performance over time as they encounter more data. The process can be categorized
into supervised learning, unsupervised learning, and reinforcement learning, each with a distinct application.
II. METHODOLOGY
In conducting this research, a critical first step was the thorough preprocessing of the dataset. Various data
preprocessing steps are performed, truth us handling missing values, dropping accessory columns, creating

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[1126]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:10/October-2023 Impact Factor- 7.868 www.irjmets.com
new features, and checking the dataset must be clean and no null values in datasets. For machine learning some
libraries are used like the Num-py used for numerical operation, Pandas used for data manipulation and
analysis, Matplotlib used for creating plots, Seaborn used for data visualization, Scikit-learn used for machine
learning tasks. Other standard libraries are used for file handling and data processing. Code loads two datasets
train matches and train deliveries which contain info about cricket matches and deliveries. Data loaded into
pandas data frames. The code includes various exploratory data analysis tasks, such as visualizing the number
of matches played by each team, analyzing toss winners and match winners, and more. For predicting the
winner we evaluate the model like a Naïve Bayes, Decision Tree Regressor, Support Vector Classifier, and
Random Forest Classifier. In model evaluation, each model is trained on the training data and evaluated on a
test set, after that accuracy scores and confusion matrices are calculated to assess model performance testing
also plays an important role in the model. The dataset (presumably for future cricket matches) is loaded and
preprocessed using the same feature engineering steps as the training data.
Machine learning
Machine learning is to predict which team will win a cricket match based on various features and historical
data. The performance of different machine learning models is assessed and compared for this prediction task.
Demonstrates these prediction models and provides a basis for selecting the most appropriate model for
predicting cricket match outcomes.
III. RESULTS AND DISCUSSION
Accuracy.
After the model training, many models have a different accuracy of each model.
Table 1. Classification report
MODEL NAME ACCURACY
NAVIE_BAYES ALGORITHM 91.0
DECISION TREE REGRESSOR 89.0
RANDOM FOREST ALGORITHM 88.0
SUPPORT VECTOR ALGORITHM 60.0

Figure 1: Naïve bayes algorithm

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[1127]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:10/October-2023 Impact Factor- 7.868 www.irjmets.com

Figure 2: Decision tree regressor

Figure 3: Random forest algorithm

Figure 4: Support vector algorithm

IV. CONCLUSION
In conclusion, the use of machine learning in predicting cricket match outcomes is a promising and valuable
approach. Cricket is a complex sport influenced by various factors such as player performance, pitch conditions,
and team dynamics, making it challenging to predict match results with conventional methods. Machine
learning leverages historical data, algorithms, and statistical models to identify patterns and relationships,
www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science
[1128]
 e-ISSN: 2582-5208
International Research Journal of Modernization in Engineering Technology and Science
( Peer-Reviewed, Open Access, Fully Refereed International Journal )
Volume:05/Issue:10/October-2023 Impact Factor- 7.868 www.irjmets.com
providing valuable insights and accurate predictions. This research highlights the potential of machine learning
in cricket match prediction and serves as a basis for selecting the most appropriate model for forecasting
cricket match outcomes. However, it’s essential to recognize that no prediction model is foolproof due to the
inherent uncertainty in sports, and various factors can influence the game’s outcome. Nevertheless, machine
learning offers a data-driven approach to enhance the accuracy of cricket match predictions, benefiting cricket
enthusiasts, analysts, bookmakers, and teams seeking data-driven insights for strategizing effectively.
V. REFERENCES
[1] M. J. Awan et al., “Cricket Match Analytics Using the Big Data Approach,” Electronics, vol. 10, no. 19, p.
2350, Sep. 2021, doi: 10.3390/electronics10192350.
[2] A. Sinha, “Application of Machine Learning in Cricket and Predictive Analytics of IPL 2020,” other,
preprint, Oct. 2020. doi: 10.20944/preprints202010.0436.v1.
[3] V. Sivaramaraju, N. Sethi, and R. Rajender, “Heuristics for Winner Prediction in International Cricket
Matches,” Stat., optim. inf. Comput., vol. 8, no. 2, pp. 602–609, May 2020, doi: 10.19139/soic-2310-
5070-648.
[4] D. M. Vistro, F. Rasheed, and L. G. David, “The Cricket Winner Prediction With Application Of Machine
Learning And Data Analytics,” vol. 8, no. 09, 2019.
[5] I. Wickramasinghe, “Applications of Machine Learning in Cricket: A systematic review,” Machine
Learning with Applications, vol. 10, p. 100435, Dec. 2022, doi: 10.1016/j.mlwa.2022.100435.

www.irjmets.com @International Research Journal of Modernization in Engineering, Technology and Science

[1129]