Diabetes treatment pattern identification through layered Tri-skip-gram approach

Abstract

It’s very much crucial to suggest a compatible treatment pattern for a disease, based on its various symptoms at different stages of it. Twitter is a powerful form of social media for sharing information about various issues and can be used to raise awareness and collect pointers about associated risk factors and preventive measures. Twitter tweets retrieved with the Hashtag '#Diabetes’ is the origin resource of my paperwork. Here I proposed a recommendation model for suggesting treatments for Diabetes considering social media’s Twitter data set which has undergone Bigram & Trigram Analysis to provide the analysis report of people suffering from diabetes at different stages and the treatments suggested for different stages differently. The technique proposed is a layered Tri-skip-gram approach which finds widespread use in the analysis of Textual data. Here the process starts with relevant diabetes tweets retrieval and then the task of normalization to get required tweet text for the textual analysis. This retrieval process ensures the presence of the unigram ‘diabetes’ in all the tweets. At the second level, Bigram classification was implemented for the status tracking of Diabetes for which the outcome is different clusters groups of diabetes tweets for different stages of it. Now the main task of this paper is classifying these cluster groups. The proposed model uses feature extraction to find treatment groups and then uses term-tweet classification method to derive the patterns of diabetes treatments suggested for different stages of diabetes as Prediabetes, Type2, and Type1.