Automatic Drug Reaction Detection Using Sentimental Analysis
L Girish · 2019
NowadaysvariousPharmaceuticalsindustries are using the Social media in whichindustries may include sick persons, Doctors,Pharmaceuticals companies etc. There exists a lot ofmedical sites and conference blogs which are the part ofsentimental analysis. These are used in manyapplications such as text based information (i.e., socialmedia such as twitter, Facebook ), market intelligence,drug observations ,taking the opinions about heathrelated information and also to detect the impact ofAdverse Drug Reaction(ADR) automatically by makinguse of Pharmacovigilance techniques. By using thealready existing tools which are used to examine thehealth related information and sentimental analysis weare not able to get the sufficient classificationaccuracies. And also they are not able to extract theexploratory and transferable features; hence they arelacking the biased features. In our design we use theAdvanced NLP approaches to generate the functionalfeatures from the input and using the advancedmachine learning algorithms to achieve theclassification accuracies. Here we rely on textclassification approach that generates huge set offeatures which represents many semantic propertieslike sentiment, polarity, topics which are useful tomanifest the experience of the user when they talkabout ADR. We present two datasets that are preparedby user to perform the job of ADR detection from thedata posted by the user on internet. We also verifywhether the combining training data which have beentaken from different collections can improve theautomatic classification accuracies and address the issueof data imbalance. This is very useful where there existsa large datasets and also reduces time and cost of thosedata.