Hybrid Recommender System for Mental Illness Detection in Social Media Using Deep Learning Techniques

Table 1

Comparison of existing methodology.

Method	Description	References
Unified relevance model	It is a probabilistic item-to-user relevance framework that uses the parzen-window approach to estimate the density of relevant items. This strategy helps to alleviate the issue of data sparsity.	Si and Jin [17]
Hybrid CF model	Effective recommender systems are introduced, which make use of sequential mixture CF and joint mixture CF to achieve their results. It also incorporates sophisticated bayes belief theory.	Su et al. [18]
Fuzzy association rules and multilevel similarity (FARAMS)	It makes use of furzy association rule mining in order to expand the capabilities of the current methodologies. It was possible for FARAMS to complete the goal of producing higher qualitative forecasts.	Wang et al. [19]
Flexible mixture model (FMM)	The formation of user and item clusters might happen at the same time. It adds preference nodes in order to investigate a significant variance in rating among users who have similar preferences.	Leung et al. [20]
Maximum entropy approach	In order to lower the apriori likelihood of an item, it is clustered depending on the user’s access route. This is beneficial in dealing with sparsity and dimensionality.	Pavlov and Pennock [21]