Emotion detection(ED) and analysis from social media data have been adopted by applications in a number of sectors, including e-commerce, healthcare, and the media. Popular social media sites for opinion sharing include LinkedIn, Weibo, Twitter, Facebook, and Weibo. People can be greatly influenced by the feedback and tweets that others share, as well as by the level of contact that is accessible on social media platforms. Selecting the essential data is incredibly tough and complex because there is an abundance of data available. This study focuses on depression, a type of emotion that has a detrimental effect on people's day-to-day lives. This study proposes a design of a computationally feasible feature selection and deep learning (DL) based classification of depression-related tweets. A benchmark Twitter emotion detection dataset is used to train and assess the suggested classifier. The findings of the experiment show that the proposed approach performs more effectively than the most advanced classical machine learning (ML)-based emotion recognition methods, attaining the highest levels of accuracy, recall, F1-score, and precision.