A NEW APPROACH FOR BREAST CANCER DETECTION - BASED MACHINE LEARNING TECHNIQUE

Aug 3, 2026

Upload by Violet

Breast cancer causes a major share of cancer-related deaths, making early detection essential. This research presents an automated breast cancer diagnosis system using a machine learning model in which convolutional neural networks (CNNs) select discriminative features and contrast limited adaptive histogram equalization (CLAHE) reduces noise. The work compares Random Forest, SVM, KNN, Naïve Bayes, and Logistic Regression on a combined dataset of 3002 images from 1400 individuals assessed by digital mammography (2007–2015). Model performance is evaluated using accuracy and precision, showing strong efficiency with low computing requirements.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

INTRODUCTION

Breast cancer background and risk factors

Breast cancer symptoms and imaging examples

Cancer subtypes and classification overview

Motivation for machine-learning based diagnosis

Submitted: 2024-01-19 | Revised: 2024-02-27 | Accepted: 2024-02-29 Keywords: machine learning, breast cancer, CNN, image processing, healthcare MalekM. AL-NAWASHI [0000-0002-5231-8155]* , Obaida M. AL-HAZAIMEH [0000-0001-5641-4892]* , Mutaz Kh. KHAZAALEH [0000-0002-2071-7020]* A NEW APPROACH FOR BREAST CANCER DETECTION-BASED MACHINE LEARNING TECHNIQUE Abstract The leading cause of cancer-related mortality is breast cancer. Breast cancer detection at an early stage is crucial. Data on breast cancer can be diagnosed using a number of different Machine learning approaches. Automated breast cancer diagnosis using a Machine Learning model is introduced in this research. Features were selected using Convolutional Neural Networks (CNNs) as a classifier model, and noise was removed using Contrast Limited Adaptive Histogram Equalization (CLAHE). On top of that, the research compares five algorithms: Random Forest, SVM, KNN, Naïve Bayes classifier, and Logistic Regression. An extensive dataset of 3002 combined images was used to test the system. The dataset included information from 1400 individuals who underwent digital mammography between 2007 and 2015. Accuracy and precision are the metrics by which the system's performance is evaluated. Due to its low computing power requirements and excellent accuracy, our suggested model is shown to be quite efficient in the simulation results. 1. INTRODUCTION Everyone, regardless of age or socioeconomic status, is vulnerable to the devastating effects of cancer. There are various varieties of cancer, but breast cancer is one of the most common tumors among women. This difficulty underscores the need for researchers to focus on improving cancer diagnosis and prognosis. Machine-learning strategies may have a major impact in cancer early detection and prognosis. Breast cancer, which begins in the ducts of the breast, affects a disproportionately high number of women globally. Among cancers that affect women, breast cancer is the second highest cause of mortality (Fatima et al., 2020; Houssein et al., 2021;Yassin et al., 2018) . Breast cancer risk may be increased by the following factors: Age is a significant determinant in breast cancer development. Although females have a higher incidence of breast cancer, males are not impervious to the condition. * Al-Balqa Applied University, Information Technology Department, Jordan, [nawashi@bau.edu.jo](nawashi@bau.edu.jo), [dr_obaida@bau.edu.jo](dr_obaida@bau.edu.jo), [mutaz.khazaaleh@bau.edu.jo](mutaz.khazaaleh@bau.edu.jo) An individual's risk is increased by having a mother, daughter or sibling who has received a breast cancer diagnosis. The susceptibility to breast cancer is elevated in individuals harboring specific gene mutations, including the genes BRCA1 and BRCA2 . Prolonged utilization of HRT (i.e., "Hormone Replacement Therapy ") has the potential to elevate the associated risk. Delayed puberty or lack of children, premature menstruation, and postponed menopause are all variables that may influence a woman's risk (Nanda et al., 2002) . Breast cancer symptoms, as shown in Figure 1, include unexplained lumps in either breast, changes in breast size, shape, or appearance, pain in either breast, nipple discharge other than breast milk, and changes in the texture or color of the breast skin (Svensson et al., 2020) . |
Lumps

Nipple Pain

Skin's Texture |
Discharge Nipple

Nipple Retraction

Lymph Node | | | --- | --- | --- | Fig. 1.Breast cancer symptoms Figure 2 displays a single example of each distinct abnormality type, including mass, carcinoma, calcification, and asymmetry (Chang et al., 2021; Svensson et al., 2020) . | | | |
Mass | | --- | --- | --- | --- | | | | Carcinoma | | Fig. 2.Various forms of anomalies The morphology of breast cancer cells is utilized to categorize the disease into various subgroups. DCIS (i.e., "Ductal Carcinoma In Situ") is a type of breast cancer

FAQs

How does the proposed breast cancer detection system preprocess data and extract features?

It uses CLAHE to remove noise and convolutional neural networks (CNNs) to select features used by the classifier.

Which machine learning algorithms are compared in the research?

The study compares Random Forest, SVM, KNN, Naïve Bayes classifier, and Logistic Regression.

What dataset is used to test the system and how is performance evaluated?

The system is tested on 3002 combined images from 1400 individuals who underwent digital mammography between 2007 and 2015, with accuracy and precision used as evaluation metrics.

Aug 3, 2026

Upload by Violet