Feature Selection using Ant Colony Optimization for Microarray Data Classification

Abstract

Microarray datasets have very high dimensions and contain noise and redundancy. The issue with the microarray dataset is that it includes more characteristics than the samples, which reduces the algorithm’s effectiveness. To put it another way, compared to rows, there are more columns. To precisely extract data from the microarray dataset, we need a powerful approach. Microarray datasets are essential for diagnosing cancer, tumours, and other disorders. So here is the feature selection strategy in action. Recently, feature selection (FS) has gained in importance as a data preparation strategy, particularly for high-dimensional data. We favour classification issues with fewer characteristics and high accuracy since we know that all attributes are not required to attain high accuracy. The main objective of feature selection is to identify the best subset of features. In order to accomplish this goal, we are utilizing the Ant Colony optimization (ACO) method, which is a technique that takes inspiration from the foraging habits of ant colonies. We will be using ACO combined with Logistic Regression (LR), Decision Tree (DT) and Random Forest (RF), and. We will compare the accuracy of every algorithmic model and every dataset, as well as fitness error. Random Forest and Logistic Regression has performed well for both datasets.