A Comparative and Predictive Analysis of Prostate Cancer Diagnosis and Treatment using Decision Tree, Neural Network, Support Vector Machine, Random Forest and K-Nearest Neighbor KNN Classification Algorithms

• Author(s): Oguoma Ikechukwu Stanley ; Uka Kanayo Kizito ; Victory Chibuike Onumaku ; Onu-Njoku Charles Enyinnaya ; Nnedinma Christiana Njoku; Chukwu Alphonsus Chekwube
• Paper ID: 1705354
• Page: 170-187
• Published Date: 13-01-2024
• Published In: Iconic Research And Engineering Journals
• Publisher: IRE Journals
• e-ISSN: 2456-8880
• Volume/Issue: Volume 7 Issue 7 January-2024

Abstract

Prostate cancer is an illness majorly found on men between age of 50 and above. It begins when the healthy cells in the prostate gland change and grow out of control, and grow a mass called tumor which might affect any part of the body whereby could become cancer cells, and then spread to other parts of the body. This study is centered on a comparative predictive analysis of prostate cancer diagnosis and treatment using Decision Tree, Neural Network, Support Vector, Random Forest and K-Nearest Neighbor KNN Classification Algorithms. The study tends to achieve the following objectives: are to design a more accurate and intelligent model for easy identification and diagnosis of prostate cancer disease on patients', to compare the accuracy of the results produced between the Five (5) algorithms in other to proffer a more long-lasting solution for prostate cancer disease prediction and hence lower mortality rate amongst patients. The research adopted a data mining methodology called classification algorithm by following the SEMMA (sample Explore modify model Access) approach while employing Five (5) machine learning algorithm as the modeling tool. The experiment on the collected prostate cancer dataset was analyzed with R Programming language while using JASP IDE for the experiment sourced from UCI machine learning repository. The result was able to implement a model that could easily and accurately predict the presence of prostate cancer in men efficiently and effectively with Decision Tree 80% test accuracy, Neural Network Algorithm 90% test accuracy, Support Vector Machine (SVM) 75% test accuracy, Random Forest Algorithm 80% test accuracy and Out-of-Bag (OOB) accuracy of 90% and K-Nearest Neighbors (KNN) algorithm 90%. Base on the comparison analysis conducted by this study, it was observed that Neural Network Algorithm and K-Nearest Neighbors (KNN) have the highest percentage accuracy towards the prediction of prostate cancer disease having 90% test accuracy each with KNN 1% validation accuracy. This research was able to show clearly how prostate cancer disease could be managed using prediction models on the tested 80% trained dataset on the various algorithms used for the experiment.

Keywords

Artificial Intelligence, Machine Learning, Prostate Cancer, Decision Tree, K-Nearest Neighbor, Vector Machine, Random Forest and Neural Network Algorithms

Citations

IRE Journals:
Oguoma Ikechukwu Stanley , Uka Kanayo Kizito , Victory Chibuike Onumaku , Onu-Njoku Charles Enyinnaya , Nnedinma Christiana Njoku; Chukwu Alphonsus Chekwube "A Comparative and Predictive Analysis of Prostate Cancer Diagnosis and Treatment using Decision Tree, Neural Network, Support Vector Machine, Random Forest and K-Nearest Neighbor KNN Classification Algorithms" Iconic Research And Engineering Journals Volume 7 Issue 7 2024 Page 170-187

IEEE:
Oguoma Ikechukwu Stanley , Uka Kanayo Kizito , Victory Chibuike Onumaku , Onu-Njoku Charles Enyinnaya , Nnedinma Christiana Njoku; Chukwu Alphonsus Chekwube "A Comparative and Predictive Analysis of Prostate Cancer Diagnosis and Treatment using Decision Tree, Neural Network, Support Vector Machine, Random Forest and K-Nearest Neighbor KNN Classification Algorithms" Iconic Research And Engineering Journals, 7(7)