Machine learning (ML) and deep learning (DL) strategies are excelling in very effective ways to solve complex as well as difficult real-world problems. This study aims to present a comparative analysis of conventional ML and DL methodologies based on various factors. This study has considered into account various kinds of generalized factors including working mechanisms, requirements and dataset reliance, accuracy, result interpretation, human participation during model training, complexity, applications, etc. After a conceptual analysis, an experimental analysis has been also done utilizing a sufficient data set to illustrate the potential of both the techniques. The present study finds that machine learning performs effectively with small and structured datasets whereas deep learning is a more successful option for complex, diverse and large data sets. The article begins with an introduction and evolution of artificial intelligence. Subsequently, an introduction to ML as well as DL and a literature review along with a research gap have been presented. Furthermore, this work presents a detailed comparative study, experimental results analysis followed by a conclusion and future perspectives.