The increased level of financial transactions before and specifically after the influx of COVID-19 has heightened the activities of fraudsters in the mobile money sector. This calls for the development of robust systems that would effectively detect and if possible prevent these unscrupulous occurrences to a large extent. The features of mobile money transactions dataset are highly unstructured, and therefore need to be streamlined using a powerful supervised machine learning tool. Genetic Algorithm has been proposed as an effective feature selection method in this paper. Selected machine learning algorithms were used to further build the models for the fraud detection system. These ML techniques include Artificial Neural Networks, Random Forest, Naïve Bayes, Decision Trees and Logistic Regression. In order to test the efficacy of the classifiers, the performance of the models were validated using mobile money dataset obtained from kaggle database. The findings from this work have clearly proven that the proposed algorithm used for the feature selection exhibits greater efficiency than the existing machine learning techniques.

Four supervised machine learning algorithms were used in building social engineering attack prevention models i.e. Support Vector machines, Logistic Regression, Bayesian Network and Random Forest. The efficacy of the models was ascertained by the help of three datasets (one synthetic and two real-life). The performances of the machine learning methods were compared using F1-score, accuracy, Matthew’s Correlation Coefficient, Precision and Sensitivity. The usefulness of each dataset was also compared per machine learning technique to draw relevant conclusions. Results from the experimental set-up showed that SVM demonstrates greater resilience in preventing mobile money fraud attacks as compared to the other algorithms. Again, even though real datasets are complex and utilize greater system resources during model training and testing, they help in building more effective social engineering attack prevention models than other datasets that are obtained from synthetic sources.