Detecting Spam Email with Machine Learning Optimized with Harris Hawks’s optimizer (HHO) Algorithm

Aim:

       This study aims to improve the accuracy of spam email detection by leveraging a hybrid approach employing the Harris Hawks Optimizer (HHO) in conjunction with the powerful XGBoost algorithm for feature selection in machine learning.

Abstract:

            The persistent and evolving threat of spam emails in today’s digital ecosystem necessitates advanced and adaptive mechanisms for their detection and filtration. In this pursuit, this research presents an innovative framework aimed at significantly augmenting the precision and efficacy of identifying spam emails. Leveraging machine learning paradigms, the study introduces a hybridized approach that integrates the Harris Hawks Optimizer (HHO) with the robust XGBoost algorithm, specifically targeting feature selection. The primary aim is to substantially elevate the accuracy and computational efficiency in the differentiation between spam and genuine emails.

Existing Method:

            The existing method employed the HHO algorithm in tandem with the K-Nearest Neighbours (KNN) algorithm for feature selection. While effective, it had limitations in achieving optimal accuracy and feature identification.

Problem Definition:

               Spam email detection is a critical challenge in today’s digital landscape. The problem lies in accurately identifying relevant features that distinguish spam from legitimate emails. The goal is to enhance detection accuracy.

Proposed Method:

                This research proposes an enhancement by replacing KNN with XGBoost, a more powerful and flexible algorithm. The HHO algorithm is employed for feature selection to optimize the feature subset for XGBoost, improving the classification accuracy and computational efficiency.