Classification of benign and malignant tumor cells using random tree and knn classifiers from Wisconsin dataset for the potential diagnostic application

• Published in: AIP conference proceedings Vol. 2816; no. 1
• Main Authors: Koyyala, Umakanth, Thirunavukkarasu, Usharani
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 22.03.2024
• Subjects: Classifiers; Data mining; Data systems; Datasets; Diagnostic software; K-nearest neighbors algorithm; Machine learning; Software; Statistical analysis; Tumors; United States > US; Wisconsin
• ISSN: 0094-243X; 1551-7616
• DOI: 10.1063/5.0186646

Aim: The study’s goal is to categorise benign and malignant tumour cells from the Wisconsin dataset using Random tree and K-nearest neighbour classifiers for possible diagnostic applications. Materials and Methods: Breast cancer report and data system (WDBC) with benign (n=21) and malignant (n=21) masses are collected from the kaggle machine learning repository for the proposed study. The dataset contains 21 attributes which are considered as inputs to our study. The classification of diseased and healthy subjects was performed using WEKA, a data mining tool. IBM SPSS software was used for the statistical analysis. The classifier’s performance was compared using an independent sample t-test and SPSS software. There was a statistically significant difference (p0.05) between the groups. The KNN classifier achieved the classification accuracy rate as 95.23%, which is higher than the Random tree classifier (90.47%). Conclusion: The KNN classifier shows the higher accuracy value when compared to the Random tree classifier for predicting the benign and malignant tumor cells using WDBC dataset.