As the technology has progressed, we have seen substantial rise in mental health issues, particularly among youths. Early detection of depression, stress, suicidal ideation ,etc. can ensure timely intervention and support for such individuals.
It labels the social media post of the user according to the possible mental health issue.
The model has been trained on Reddit posts, where the subreddit serves as the label for the text. We have employed LSTM model for multiclass classification.
The main challenge was the accuracy and finetuning of the model and ensuring proper preprocessing of dataset without losing important information.
The model has good accuracy for a multiclass classification
Hybrid models can perform better. The model can be further improved and has real world impact.
Making it versatile as well as accurate is the main goal ahead.
Log in or sign up for Devpost to join the conversation.