Simple Machine Learning Architecture as a Service

Case Study: Gender Prediction

Abstract

Machine learning (ML) development starting in the 1950s, has shown significant progress. Various fields have used machine learning as an information system element that is useful in assisting data processing, personalization, prediction, and performing anomaly detection of occurring transactions. Along with developments, cloud-based machine learning technology is becoming the choice for ease of implementation and connectivity with various other technology platforms. This paper proposes a machine learning architecture as a service (MLaaS) implemented in a case study of a gender prediction model based on height and weight. The results show that the MLaaS architecture is straightforward to implement and fits the needs of various access environments and the ease of updating models centrally. Our gender prediction model achieved 91.78% in the precision, recall, and F1-score, 91.8% in specificity and NPV, and 91.79% in accuracy.