INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Development of a Predictive Model for Prostate Cancer Using a

Machine Learning Based Classification Algorithm

Akinrolabu, Olatunde David

Deep-Sight Research Group, Department of Computer Science, Adekunle Ajasin University, Akungba-

Akoko, Ondo State, Nigeria

DOI : https://doi.org/10.51583/IJ L T EMAS.2025.1411000108

Received: 08 December 2025; Accepted: 15 December 2025; Published: 23 December 2025

ABSTRACT

Prostate cancer remains one of the most prevalent malignancies among men worldwide, with early detection being

crucial for effective treatment and improved survival outcomes. Traditional diagnostic procedures, such as

prostate-specific antigen (PSA) testing, digital rectal examination (DRE), and biopsy, often suffer from limitations

including subjectivity, low specificity, and inconsistent accuracy. This study presents the development of a

predictive model for prostate cancer using a machine learning-based classification algorithm, specifically the

Support Vector Machine (SVM). The dataset utilised was obtained from a publicly available prostate cancer

repository, containing relevant biomedical and demographic features. Preprocessing procedures, including

normalisation and data transformation, were applied to enhance model quality and ensure robustness.

Experimental results revealed that the SVM model achieved a high predictive accuracy of 84.8% under

crossvalidation and 87% on full dataset evaluation, with a corresponding error rate of less than 0.17. These results

demonstrate the model’s ability to accurately distinguish between malignant and non-malignant cases, validating

its suitability for clinical decision support. The model’s performance further confirms the potential of SVM as an

effective classification technique for medical diagnostics, especially where datasets exhibit complex, nonlinear

feature interactions. This study emphasises the significance of machine learning in enhancing diagnostic precision

and reliability within the medical domain. The outcomes provide valuable insights into integrating artificial

intelligence into healthcare systems for early cancer detection, reducing diagnostic delays, and supporting medical

professionals in clinical decision-making.

Keywords: prostate cancer, support vector machine, machine learning, performance evaluation, classification

INTRODUCTION

Prostate cancer is one of the most prevalent malignancies among men and remains the second leading cause of

cancer-related deaths worldwide (Zhang & Xiang, 2013). Globally, its incidence exhibits significant geographical

variation, with the highest rates reported in the United States, Canada, and Scandinavian countries, while the

lowest are found in Asian populations, particularly in China. In Nigeria, prostate cancer poses an increasingly

serious public health challenge. Mohammed et al. (2020) reported approximately 161,360 diagnosed cases in

2017, resulting in an estimated 56,730 deaths. The risk of developing prostate cancer increases with age and is

influenced by several factors such as genetic predisposition, ethnicity (notably higher among men of African

descent), diet, and family history. Early detection of prostate cancer is crucial for improving survival outcomes,

as timely intervention can enhance the five-year survival rate for up to nine out of ten patients (Abdel-Zaher &

Eldeib, 2018). However, early detection remains a major clinical challenge due to the limitations of existing

diagnostic methods. Traditional techniques such as the digital rectal examination (DRE), prostatespecific antigen

(PSA) testing, transrectal ultrasound (TRUS), and biopsy procedures, although widely adopted, often yield

inconsistent or inconclusive results. PSA testing, for instance, has been shown to reduce mortality by about 20%,

yet it is associated with issues of overdiagnosis and overtreatment, as it cannot accurately predict tumour

aggressiveness. Similarly, TRUS-guided biopsy lacks sufficient sensitivity to detect all clinically significant cases

(Schröder et al., 2018).

www.ijltemas.in

Page 1142

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

With advancements in medical imaging, multiparametric magnetic resonance imaging (MRI) has emerged as a

more reliable diagnostic tool, offering improved sensitivity and specificity compared to TRUS, particularly for

lesions in the transition zone (Abdelhamied, 2018). Nevertheless, despite these technological developments,

challenges persist in differentiating between aggressive and indolent prostate cancers, emphasising the need for

intelligent computational systems capable of enhancing diagnostic accuracy and risk stratification. In Nigeria, the

burden of prostate cancer is further compounded by systemic health sector challenges, including inadequate

diagnostic infrastructure, a shortage of trained medical personnel, limited access to advanced healthcare

technologies, and insufficient data management systems. These factors collectively hinder timely detection and

effective treatment, leading to higher morbidity and mortality rates. Addressing these limitations requires

innovative and data-driven approaches capable of supporting medical decision-making and improving clinical

outcomes. Motivated by these challenges, this study seeks to contribute to the advancement of intelligent

healthcare systems through the development of a predictive model for prostate cancer detection using machine

learning techniques. Specifically, this research aims to design a model capable of accurately classifying prostate

cancer risk levels based on relevant clinical features to show how computational intelligence can enhance

decision-making in prostate cancer management and boost early detection by combining machine learning with

clinical diagnostic data, especially in healthcare systems with low resources like Nigeria's.

LITERATURE REVIEW

In recent years, computational intelligence has emerged as a transformative approach for addressing complex

problems in various domains, including healthcare. Techniques such as fuzzy logic, expert systems, neural

networks, and machine learning algorithms have been widely applied to medical diagnostics, demonstrating

promising results in decision support and disease classification (Samy & Naser, 2018). technologies aim to reduce

diagnostic errors, improve efficiency, and enhance accessibility to quality healthcare, particularly in developing

countries where medical expertise and infrastructure are limited.

A. Expert Systems and Fuzzy Logic in Medical Diagnosis

Expert systems have played a vital role in medical diagnostics by simulating the reasoning capabilities of human

experts. Samy and Naser (2018) developed an expert system using the CLIPS programming language to assist in

the analysis of cancer-related conditions. The system demonstrated encouraging preliminary results, receiving

positive feedback from users. Similarly, Abdelhamied et al. (2021) designed an expert system to assist healthcare

workers in overcrowded Egyptian outpatient clinics by encoding medical knowledge of over 300 common

diseases into a rule-based system. The system utilised production rules triggered by specific symptom

combinations to provide diagnostic hypotheses and treatment recommendations. Although effective, the system

faced challenges such as selection bias due to specific imaging requirements and underfitting during model

testing.

Fuzzy logic has also been applied successfully in medical diagnosis, offering a mechanism for reasoning under

uncertainty. Mohammed (2018) developed a fuzzy expert system for diagnosing back pain diseases using

parameters such as body mass index, age, gender, and physical symptoms. The system achieved 90% diagnostic

accuracy, highlighting the potential of fuzzy logic in handling vague or imprecise medical data. These

developments underscore the importance of hybrid intelligent systems in clinical decision-making, although their

rule-based structures often limit scalability and adaptability to larger, more diverse datasets.

B. Machine Learning and Deep Learning Applications

Machine learning (ML) has gained considerable attention for its ability to learn complex patterns from medical

data without explicit programming. Ertosun and Rubin (2020) employed three architectures of convolutional

neural networks (CNNs) to identify malignant masses in mammography images from the DDSM dataset. Their

data augmentation techniques cropping, translation, rotation, and scaling—improved model generalisation,

although CNN-based methods require extensive computational resources and large datasets. Abdel-Zaher and

Eldeib (2020) proposed a deep learning framework that initialised neural network parameters using a pre-trained

deep belief network (DBN). The approach demonstrated enhanced classification performance but was limited by

www.ijltemas.in

Page 1143

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

the curse of dimensionality, emphasizing the need for efficient dimensionality reduction strategies in medical

image analysis.

In another study, Priya et al. (2019) proposed a colour-based image segmentation framework for breast

thermography analysis using the k-means clustering algorithm. Their model effectively detected abnormal thermal

patterns indicative of tumours. However, k-means clustering faced limitations such as difficulty in predicting the

optimal number of clusters (k), inconsistent results across runs, and poor performance with nonglobular cluster

distributions. Sadhana et al. (2021) compared several supervised learning classifiers, including Decision Trees

and Support Vector Machines (SVMs), on three breast cancer datasets—Wisconsin Breast Cancer (WBC),

Wisconsin Diagnostic Breast Cancer (WDBC), and Wisconsin Prognostic Breast Cancer (WPBC). The SVM

achieved the highest classification accuracy, reinforcing its robustness and reliability in biomedical data

classification tasks.

C. Hybrid and Knowledge-Based Diagnostic Systems

Several researchers have explored the integration of rule-based reasoning with learning algorithms to enhance

diagnostic precision. Roventa et al. (2022) developed an expert system designed to identify major cancerous

diseases using clinical and paraclinical data. The system contained a structured knowledge base encompassing 27

diseases across nine categories, thereby improving the diagnostic reasoning process for cases with overlapping

symptoms. Nonetheless, the study’s scope was limited to a small patient population, restricting its generalizability.

Similarly, Qeethara-Kadhim et al. (2018) evaluated the effectiveness of artificial neural networks (ANNs) in

disease diagnosis, applying feed-forward backpropagation networks to detect acute nephritis and heart disease.

The model achieved impressive classification accuracies of 99% and 95% respectively, demonstrating the

potential of ANNs in clinical classification problems. However, such models often face challenges related to

overfitting, explainability, and the need for large, annotated datasets to ensure robust performance across diverse

patient populations.

Freasier et al. (2019) constructed a medical expert system using a multiplication model of support logic

programming to determine the dominant stenosis in coronary arteries based on preprocessed myocardial perfusion

images. Using a Prologue-based knowledge base, the system correctly identified the location of arterial stenosis

in over 90% of cases, underscoring the efficiency of hybrid symbolic–statistical systems in diagnostic inference.

3.4 Summary and Research Gap

From the reviewed literature, it is evident that computational intelligence and machine learning techniques have

significantly contributed to the automation of medical diagnosis across various disease domains. However, most

existing studies either focus on specific cancers, such as breast or skin cancer, or rely on small datasets that limit

model generalisation. Moreover, few studies have explored the application of advanced machine learning

techniques, particularly Support Vector Machines (SVMs), in prostate cancer prediction within low-resource

healthcare contexts.

Problem Statement

Despite the growing global attention on prostate cancer, effective early detection remains a significant challenge,

particularly in developing nations such as Nigeria. The conventional diagnostic process for prostate cancer heavily

relies on the availability of qualified specialists and the use of standard clinical procedures, including PSAtesting,

digital rectal examination, and biopsy. However, these methods are often limited by subjectivity, high false-

positive rates, and dependence on expert interpretation. In many healthcare facilities, especially in resource-

constrained settings, the absence of specialized oncologists or radiologists further delays diagnosis and treatment,

resulting in poor patient outcomes. Moreover, manual diagnostic procedures are timeconsuming, prone to human

error, and difficult to scale in regions with limited healthcare infrastructure. These systemic challenges highlight

the need for intelligent, data-driven diagnostic systems that can assist medical practitioners by providing accurate

and rapid prediction of prostate cancer risk. Such systems could serve as complementary diagnostic tools,

particularly in settings where access to medical expertise is restricted. Therefore, the core problem addressed in

www.ijltemas.in

Page 1144

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

this study is the lack of an automated, machine learning–based predictive framework capable of supporting early

and accurate detection of prostate cancer. Developing such a computational model has the potential to enhance

diagnostic precision, reduce dependence on specialist availability, and contribute to more efficient and equitable

healthcare delivery.

RESEARCH METHODOLOGY

This study adopts a quantitative research approach, leveraging computational intelligence techniques to develop

a predictive model for prostate cancer diagnosis. Quantitative methods are particularly suited for data-driven

investigations, as they enable statistical evaluation of model performance and reproducibility of results. The

research process involves four key phases: data collection, preprocessing, model development, and performance

evaluation. This study adopts a support vector machine supervised learning approach. For the aim of this research

work to be achieved, the following procedures/processes shall be done to achieve the aforementioned specific

objectives.

1.

2.

3.

To collect and preprocess prostate cancer datasets containing clinically relevant features from diverse

patient samples;

To design a predictive model employing the Support Vector Machine (SVM) algorithm for accurate

classification of prostate cancer, and

To evaluate the model’s performance using standard machine learning metrics such as accuracy, precision,

recall, F-measure, and error rate.

Architecture

Figure 1: System Architecture

www.ijltemas.in

Page 1145

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Detailed Analysis of the Architecture

Data Collection

The dataset used in this study was obtained from an open-access repository, specifically curated for prostate cancer

classification tasks. The dataset comprises clinically relevant features, including demographic, genetic, and

diagnostic parameters, which were manually imported into Microsoft Excel and subsequently processed in

Python. Each record in the dataset represents an individual patient sample with corresponding feature attributes

and diagnosis labels. The data were prepared for analysis and then fed into a Support Vector Machine (SVM)

classifier to develop a model capable of distinguishing between benign and malignant prostate conditions. The

quantitative approach was chosen to facilitate measurable and objective analysis of the model’s predictive

accuracy, enabling empirical evaluation through standard performance metrics.

Figure 2: Data downloaded directly from an open-access repository

Data Preprocessing: Data preprocessing was a crucial step in ensuring data quality, consistency, and suitability

for machine learning analysis. The preprocessing pipeline involved several operations aimed at minimising noise

and improving model reliability. The key steps are outlined as follows:

a) Feature Scaling: After the label encoding, that is, all the texts are converted into numerical values, feature

scaling is necessary to normalise the range of independent variables.

The following are the steps involved in data pre-processing:

1. Read the data

2. Derive the class labels for each sample

3. Check out the missing values

4. Convert the Categorical Values

5. Split the dataset into the Training and Test Set

Model Development (Support Vector Machine)

Model Development: The Support Vector Machine (SVM) algorithm was employed for the predictive modelling

phase. SVM is a robust supervised learning algorithm that constructs optimal hyperplanes for classification in

www.ijltemas.in

Page 1146

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

high-dimensional feature spaces. It is particularly effective in medical diagnosis tasks where class boundaries are

nonlinear and datasets are moderately sized. The dataset was divided into training and testing subsets using an

80:20 ratio. The training set was used to optimise the model parameters, while the testing set evaluated the model’s

generalisation capability. Kernel functions were explored to enhance non-linear classification performance, and

hyperparameter tuning was applied to achieve optimal model accuracy. The mathematical model for the proposed

system, as adopted from Josepth and Fonten is as follows:

The SVM algorithm builds a binary classifier by constructing a hyperplane, separating class members from

nonmembers in the input space and finally finds a nonlinear decision function in the input space by mapping the

data into a higher-dimensional feature space and separating by means of a maximum margin hyperplane. The

system automatically identifies a subset of informative points called support vectors and uses them to represent

the separating hyperplane, which is essentially a linear combination of these points. This machine is presented

with a set of training examples, (x_i, y_i), where the x_iare the real-world data instances and the y_iare the labels

indicating which class the instance belongs to.

Minimize ½||W||2

(1)

(2)

Subject to Dii (WTXi - ) ≥1,i=1,…….,I

where Dii are the class labels

The parameter C is a regularisation parameter that controls the trade-off between the two terms in the objective

function. The following decision rule is used to correctly predict the class of a new instance with a minimum

error. The dual formulation permits an efficient learning of non–linear SVM separators, by introducing kernel

functions which calculate a dot product between two vectors that have been nonlinearly mapped into a

highdimensional feature space. Since there is no need to perform this mapping explicitly, the training is still

Ƒ(x)=sgn[ WT X-Y ]

(5)

The real feature space can be very high or even infinite. The parameters are obtained by solving the following

nonlinear SVM dual formulation (in Matrix form),

Minimise LD(U)=1/2 uTQu – et u

(6)

By performing computations in the input space. The decision function in this nonlinear case is given by:

Ƒ(x)=sgn[(K(x_i* T) * u – y

(7)

(8)

where u is the Lagrangian multiplier.

Minimize ½ ∑ⁿ_k=1(w^t_k)w_k+ C ∑_i¹=Eⁱ

Subject to the constraints:

Wt wi(xi) - Wtt(xi) ≥ etk - £I * k where k ≠ ki

where k_iis the class to which the training data x_ibelong,

etk = 1 – ctk

(9)

(10)

(11)

1if ki = k

c^t_k= 1(0if ki ≠ k)

The decision function for a new input data x_igiven by

www.ijltemas.in

Page 1147

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

(12)

J_i= arg max (Ƒ_x(x_i))

Ƒk(xi) = Wkt(xi) - Yk

(13)

Model Training and Testing

Model training represents the learning phase in which the predictive framework iteratively adjusts its internal

parameters to minimise classification error. During this process, the Support Vector Classifier (SVC) model is

exposed to input features along with their corresponding target labels, allowing it to learn the underlying patterns

and relationships between predictors and prostate cancer outcomes. The training process continues until

convergence is achieved that is, when further iterations yield minimal improvement in the loss function, ensuring

an optimal balance between bias and variance. Following the training phase, model evaluation is conducted using

an independent test dataset to assess its generalisation capability on unseen data. The test set provides an unbiased

estimate of the model’s predictive performance based on standard evaluation metrics. To prevent data leakage

and ensure model robustness, no instance from the training set is included in the test set. In this study, due to the

moderate size of the available dataset, the data were partitioned into two subsets: 80% (937 samples) for training

and 20% (313 samples) for testing. This stratified split ensures that both subsets maintain similar class

distributions, thereby enhancing the reliability of model performance assessment.

Figure 3: Dataset Split

Model Evaluation and Result

This is the last phase of the experimental implementation. The developed SVM model was evaluated using

standard machine learning performance metrics, including: Accuracy, Precision, Recall (Sensitivity), F1-Score

and Error Rate

These metrics collectively assess the model’s ability to distinguish between malignant and benign cases, ensuring

clinical reliability and robustness.

Figure 3: Accuracy Score, Classification Report (Precision, Recall and F-Measure)

The error generated for the epoch is calculated by taking the sum of the false negatives and false positives of the

confusion matrix (misclassified samples) and dividing it by the test data. Therefore, to calculate the total error

generated by the model the calculation is done below:

www.ijltemas.in

Page 1148

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

Error Rate(Full Set) =

= 0.1246 = 1.24%

Table 1: Evaluation Metrics (SVC Full Epoch)

Model

SVM

Accuracy Score

88%

Precision

87.5%

Recall

87%

F1-Score

87%

Error Rate

1.24%

Cross Validation

Figure 8: Cross Validation Scores

Table 2: Evaluation Metrics (Cross Validation Scores)

Epoch

Accuracy Score

80%

MEAN ACCURACY SCORE

1

2

3

4

5

6

7

8

9

10

84.2%

84%

74%

88%

81%

88%

86%

88%

80%

89%

DISCUSSION

The Support Vector Classifier (SVC) model was trained and evaluated using the preprocessed prostate cancer

dataset to predict the likelihood of cancer occurrence based on the extracted clinical and demographic attributes.

The model’s objective was to effectively distinguish between positive (cancer) and negative (non-cancer) cases

by learning the optimal decision boundary that maximises class separation within the feature space. During

experimentation, the dataset was partitioned using an 80:20 train-test split, and cross-validation techniques were

employed to assess the model’s stability and generalisation performance. The SVC demonstrated consistent

www.ijltemas.in

Page 1149

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

predictive accuracy across both evaluation settings. Specifically, the model achieved an average classification

accuracy of 84.8% during cross-validation, corresponding to a mean error rate of 0.162, indicating minimal

variance across folds. When evaluated on the complete test set, the model’s accuracy further improved to 87%,

with only 39 misclassifications out of 313 total test samples. These results confirm that the SVC effectively

captured the discriminative patterns within the data, achieving strong generalisation capability despite the

relatively limited sample size. The low error rate demonstrates the model’s robustness and ability to minimise

both false positives and false negatives, a critical factor in clinical applications where diagnostic precision directly

influences treatment outcomes. The observed performance validates the suitability of Support Vector Machines

for medical diagnosis tasks, particularly in conditions where the dataset exhibits non-linear relationships and

limited dimensionality. Furthermore, the findings suggest that integrating more diverse patient data and additional

clinical biomarkers could further enhance predictive accuracy and support broader deployment of the model as a

clinical decision-support tool for early prostate cancer detection.

CONCLUSION, CONTRIBUTION AND RECOMMENDATION

The development and evaluation of a predictive model for prostate cancer using a Support Vector Machine (SVM)

classifier yielded promising results. The study demonstrated that the model effectively classified input data with

an accuracy exceeding 84%, supported by strong precision, recall, and F1-score metrics. The model’s

performance across both train-test split and cross-validation techniques confirmed its robustness, reliability, and

minimal error convergence. These findings highlight the potential of machine learning-based systems, particularly

SVMs, to support medical practitioners in the early detection and diagnosis of prostate cancer, especially in

healthcare environments with limited specialist availability. The results also reaffirm the importance of data

quality and preprocessing in achieving high prediction accuracy and generalizable model performance.

This research contributes significantly to the growing field of computational intelligence in healthcare by

establishing the efficacy of SVM as a dependable classification approach for medical diagnosis. It further

introduces a novel Python-based implementation framework for prostate cancer prediction, offering

reproducibility and flexibility beyond traditional platforms such as MATLAB, WEKA, and RapidMiner. The

study reinforces the value of supervised machine learning models in identifying complex, nonlinear relationships

among biomedical features, demonstrating how iterative optimisation of model parameters minimises

classification error and enhances predictive accuracy. Thus, the study extends empirical knowledge on the

application of SVMs to oncological datasets and provides a foundation for future research on intelligent diagnostic

systems.

In light of these findings, future studies should focus on expanding the dataset to include more diverse and

clinically validated samples, as larger datasets often yield better generalisation and improved model robustness.

Further exploration of advanced algorithms such as ensemble learning, deep neural networks, and reinforcement

learning could enhance predictive performance.

REFERENCES

1. E. Zhang and X. Xiang, “Inference-based Naïve Bayes: Turning Naïve Bayes cost-sensitive,” IEEE

Transactions on Knowledge and Data Engineering, vol. 25, pp. 2302–2314, 2013, doi:

10.1109/TKDE.2010.49.

2. R. C. Huang and C. Lin, “Generalized Bradley–Terry models and multi-class probability estimates,”

Journal of Machine Learning Research, vol. 7, pp. 85–115, 2021. [Online]. Available:

http://portal.acm.org/citation.cfm?id=1248551

3. E. T. Mohammed, “Internet traffic classification by aggregating correlated Naïve Bayes predictions,”

IEEE Transactions on Information Forensics and Security, pp. 5–15, 2020. [Online]. Available:

http://dblp.uni-trier.de/db/journals/tifs/tifs8.html

4. O. K. Akinsola, H. Wookjoo, and J. Beom, “On the effectiveness of discretizing quantitative attributes in

linear classifiers,” Journal of Machine Learning Research, vol. 1, pp. 1–28, 2017.

www.ijltemas.in

Page 1150

INTERNATIONAL JOURNAL OF LATEST TECHNOLOGY IN ENGINEERING,

MANAGEMENT & APPLIED SCIENCE (IJLTEMAS)

ISSN 2278-2540 | DOI: 10.51583/IJLTEMAS | Volume XIV, Issue XI, November 2025

5. H. Canfield, J. Zhang, J. C. Yang, N. Milton, and A. D. Alcántara, “An explanatory analysis of driver

injury severity in rear-end crashes using a decision table/Naïve Bayes (DTNB) hybrid classifier,” Accident

Analysis and Prevention, vol. 90, pp. 95–107, 2019, doi: 10.1016/j.aap.2016.02.002.

6. Abdel-Zaher and A. Eldeib, “Evolutionary computational methods for optimizing the classification of sea

stars in cancerous diseases,” in Proc. IEEE Winter Conf. Appl. Comput. Vis. Workshops (WACVW), 2015,

pp. 44–50, doi: 10.1109/WACVW.2015.9.

7. J. Abdelhamied, “Multi-feature prostate cancer subset selection with K-neighbours classifier,”

Knowledge-Based Systems, vol. 55, pp. 110–127, 2018, doi: 10.1016/j.knosys.2013.10.019.

8. J. A. Bermejo, “Speeding up incremental wrapper feature subset selection with Naïve Bayes classifier,”

Knowledge-Based Systems, vol. 55, pp. 140–147, 2020, doi: 10.1016/j.knosys.2013.10.016.

9. Mahammad and A. Roventa, “Computerized radiographic mass detection,” Knowledge-Based Systems,

pp. 140–147, 2019, doi: 10.1016/j.knosys.2013.10.016.

10. S. Samy and N. Naser, “Prostate-specific antigen–based prostate cancer screening: Reduction of prostate

cancer mortality after correction for nonattendance and contamination in the Rotterdam Section of the

European Randomised Study of Screening for Prostate Cancer,” European Urology, vol. 65, pp. 329–336,

2014.

11. Klein, “A guide for clinicians in the evaluation of emerging molecular diagnostics for newly diagnosed

prostate cancer,” Reviews in Urology, vol. 16, no. 4, pp. 172–180, 2014.

12. J. Wein, L. R. Kavoussi, A. W. Partin, and C. A. Peters, Urology, 10th ed. Philadelphia, PA: Elsevier, 2012.

13. F. Priya, “Vitamin E and the risk of prostate cancer: The Selenium and Vitamin E Cancer Prevention Trial

(SELECT),” JAMA, vol. 306, no. 14, pp. 549–556, 2011.

14. J. Simon, “The effect of dietary and exercise interventions on body weight in prostate cancer patients: A

systematic review,” Nutrition and Cancer, vol. 67, no. 1, pp. 43–60, 2020.

15. O. J. Osisanwo and I. Thompson, “The influence of finasteride on the development of prostate cancer,”

New England Journal of Medicine, vol. 349, pp. 215–224, 2019.

16. K. Hernández, “Contemporary evaluation of the D’Amico risk classification of prostate cancer,” Urology,

vol. 70, no. 5, pp. 931–935, 2007.

17. H. Wookjoo and J. Beom, “On the effectiveness of machine learning technique attributes in linear

classifiers,” Journal of Machine Learning Research, vol. 1, pp. 1–28, 2018.

18. T. Ertosun and Y. F. Rubin, “Global cancer statistics, 2012,” CA: A Cancer Journal for Clinicians, vol. 65,

no. 2, pp. 87–108, 2020.

19. L. P. Sadhana, C. H. Bangma, and G. Leenders, “Prostate-specific antigen–based prostate cancer

screening: Reduction of prostate cancer mortality after correction for nonattendance and contamination in

the Rotterdam Section of the European Randomized Study of Screening for Prostate Cancer,” European

Urology, vol. 65, pp. 329–336, 2014.

20. B. Roventa, M. J. Alvarez, L. J. Martinez, and M. Saiz, “Prognostic role of genetic biomarkers in clinical

progression of prostate cancer,” Experimental and Molecular Medicine, vol. 47, p. e176, 2022, doi:

10.1038/emm.2015.43.

21. S. E. Qeethara-Kadhim, A. S. Kibel, and M. J. Kemeter, “A guide for clinicians in the evaluation of

emerging molecular diagnostics for newly diagnosed prostate cancer,” Reviews in Urology, vol. 16, no. 4,

pp. 172–180, 2018.

22. J. Freasier, A. Heindenreich, P. J. Bastian, J. Bellmunt, and M. Bolla, “EAU guidelines on prostate cancer.

Part 1: Screening, diagnosis, and local treatment with curative intent—update 2013,” European Urology,

vol. 65, no. 1, pp. 124–137, 2019.

www.ijltemas.in

Page 1151