The increasing integration of technologies into our daily activities offers a great opportunity to enhance individual quality of life. However, it can also have negative effects, which can lead to mental health issues and raise the risk of suicide. In fact, mental health issues among adolescents and young adults have risen significantly in recent years. Therefore, the early detection of these issues is crucial to reduce the risk of suicide. Digital media platforms offer an anonymous space, allowing users to express themselves on various topics, including their mental health status. This paper investigates the use of machine learning and deep learning models to predict mental health states and identify individuals with suicide ideation. For this purpose, we propose analyzing digital media content, such as posts and tweets, collected from two distinct datasets: SuicideIdeation Tweet and SuicideDetection. The best results were achieved by combining RNN and LSTM models in a hybrid RNN-LSTM architecture, yielding accuracy rates of 91.34% and 92.04%, with AUC values of 95.26% and 95.94% for the SuicideIdeation Tweet and SuicideDetection datasets, respectively.