Engineering structures, including civil infrastructures, have always been susceptible to various kind of damage during their service life. The goal of structural health monitoring is to provide sufficient when the structure condition deteriorates. It has the obvious benefits of preventing disastrous structural collapses and reducing maintenance costs. By and large, structural health monitoring approaches are divided into two classes, model-based, and data-driven approaches. The main challenge in data-driven approaches lies in a large amount of data to be dealt with. As machine learning algorithms are often used to recognize the inherent pattern in data, their applications in data-driven approaches have been gained an increasing attention in recent years. Utilizing machine learning algorithms turns the decision-making step of structural health monitoring into an automated process with minimal human intervention. This paper presents a summary of a literature review concerning different data-driven structural health monitoring approaches combined with machine learning algorithms developed in the last several years. Specifically, this report will review existing applications of machine learning algorithms for the purposes of dimensionality reduction and developing a statistical model for structural health monitoring. The primary aim is to categorise the existing studies in the aforementioned areas in terms of the type of machine learning algorithms. This paper also attempts to identify research gaps to facilitate the formulation of a future study in the aforementioned area.