Improving KNN by Gases Brownian Motion Optimization Algorithm to Breast Cancer Detection

Abstract

In the last decade, the application of information technology and artificial intelligence algorithms are widely developed in collecting information of cancer patients and detecting them based on proposing various detection algorithms. The K-Nearest-Neighbor classification algorithm (KNN) is one of the most popular of detection algorithms, which has two challenges in determining the value of k and the volume of computations proportional to the size of the data and sample selected for training. In this paper, the Gaussian Brownian Motion Optimization (GBMO) algorithm is utilized for improving the KNN performance to breast cancer detection. To achieve to this aim, each gas molecule contains the information such as a selected subset of features to apply the KNN and k value. The GBMO has lower time-complexity order than other algorithms and has also been observed to perform better than other optimization algorithms in other applications. The algorithm and three well-known meta-heuristic algorithms such as Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Imperialist Competitive Algorithm (ICA) have been implemented on five benchmark functions and compared the obtained results. The GBMO+KNN performed on three benchmark datasets of breast cancer from UCI and the obtained results are compared with other existing cancer detection algorithms. These comparisons show significantly improves this classification accuracy with the proposed detection algorithm.

Breast Cancer Classification based on Artificial Intelligence: State of the Art

“…In practice, the classification was performed using three UCI benchmark breast cancer datasets, and the results were compared to those obtained by other breast cancer detection algorithms. The experimental results reported in [80] validated GBMO-KNN's efficacy.…”

“…As demonstrated in [80], a new method has been developed to improve KNN's accuracy in the detection of breast cancer. Gaussian Brownian Motion Optimization with Kernel Neural Networks (GBMO-KNN) describes the proposed approach.…”

A New Fuzzy Bio-Inspired Based Classification to Cancer Detection

“…The FCOA-HPC achievement of WDBC is better than the others, execpt TVPSO-SVM and TVPSO-SVM [34]. For the hybridization of 4 hyper-planes and FCOA, there is only one competitor in the Breast Cancer Coimbra dataset, and that is a training phase of GBMO-KNN [24]. Also, CRO-HPC [5] is the only competitor of the FCOA-HPC in the training phase for the Lung Cancer dataset.…”