SPFHN: SPINAL PYRAMID FORWARD HARMONIC NETWORK FOR MENTAL HEALTH PREDICTION

Mental illness is a health issue, which undoubtedly impacts the reasoning, social interaction, and emotions of a person. These problems show the seriousness of the mental illness and the importance of early detection. Earlier detection of mental health concerns amongst people is supreme for earlier treatment as well as diagnosis by medical experts. The enhancement of health issues as well as the requirement for effectual medical healthcare has directed to research of deep learning (DL), which may be applied to mental health issues. DL modules have been introduced for predicting academic performance, yet it turns out to be complex to extract high-quality handcrafted features because of the large database. To conquer this gap, a potent approach for mental health prediction is introduced named Spinal Pyramid Forward Harmonic Network (SPFHN). First, an input text sentence is given to extract the features, which are carried out in the feature extraction phase. Then, the feature vectors are forwarded to sarcasm_state identification which is classified into sarcasm and non-sarcasm, where this process is performed by employing sarcasm_state detection algorithm. Lastly, mental health is predicted by the hybrid network SPFHN, which is incorporated by SpinalNet and PyramidNet. The evaluation metrics namely, accuracy, sensitivity and specificity achieved 92.9%, 91.5% and 93.1%, respectively.