Open-source software can be defined as software having its source code publicly accessible; either it is a web application, desktop or smartphone application, or any other type of software. Open-source software development motives are different for different types of developers. The incentives for the software development groups can vary for the adoption of this option. In software development companies, we usually come across the source code with unknown developers who need to be investigated before incorporating it into the current project. The traceback of the source code to its actual developer is much necessary. We have different techniques for this purpose. However, most of the technique does not adequately trackback due to inaccuracy. In this research, we can propose a new machine learning-based model for source code similarity detection. The proposed approach is used Latent Semantic Analysis (LSA) for similarity detection. The proposed model is justified by comparing it with state-of-the-art existing approaches using accuracy, precision, and recall measures.