Medicinal Side-Effect Analysis Using Twitter Feed
As the use of social media network has been increasing, people tend to share health-related information on social sites. Twitter is used by large number of users and it is a wide source of information to analyze the drug related side effect. In this paper
- PDF / 214,074 Bytes
- 11 Pages / 439.37 x 666.142 pts Page_size
- 44 Downloads / 185 Views
Abstract As the use of social media network has been increasing, people tend to share health-related information on social sites. Twitter is used by large number of users and it is a wide source of information to analyze the drug related side effect. In this paper, we have developed an approach to analyze the contents of tweets to identify the adverse effects of a drug. An annotated dataset is used to train SVM classifier to identify the tweets showcasing medicinal side effects. The use of feature selection and dimensionality reduction techniques have allowed us to enhance the performance of the classifier in terms of accuracy by 10.34% as well as efficiency by nearly 66.31% as compared to the previous similar approaches. Keywords Twitter
⋅
Adverse effects
⋅
Chi-square
⋅
Information gain
⋅
PCA
1 Introduction Social media networks can serve as a source to analyze individual interests and their effects on personal life as they provide huge amount of data. Out of all the social networks, Twitter has a large number of users sharing different information. Twitter allows user to become a source of knowledge and expertise about certain topic. According to statistics, 500 million tweets are generated everyday and 200 billion tweets are generated every year. This statistics show that twitter contains large amount of data which can be used as a source of analysis. P.S. Mane (✉) ⋅ M.S. Patwardhan Department of Computer Science, Vishwakrma Institute of Technology, Pune, India e-mail: [email protected] M.S. Patwardhan e-mail: [email protected] A.V. Divekar Department of Electrical Science, Savitribai Phule Pune University, Pune, India e-mail: [email protected] © Springer Nature Singapore Pte Ltd. 2018 P.K. Sa et al. (eds.), Progress in Intelligent Computing Techniques: Theory, Practice, and Applications, Advances in Intelligent Systems and Computing 719, DOI 10.1007/978-981-10-3376-6_7
59
60
P.S. Mane et al.
There are traditional approaches that use national reporting tools engaging patients, practitioners, and researchers. These allow patients to voluntarily submit the reports to about the medicinal errors. The national reporting tool such as IHI Trigger tool [Institute for Healthcare Improvement] measures adverse drug events and allows conducting a review of patient records using trigger to identify possible side effects. In United States, the voluntary reporting system such as Patient Safety Network (PSNET) has been developed by U.S. Department of Health and Human services [1] that allows healthcare professionals to submit the cases that highlight medical errors and adverse effects. As the use of social networks is emerging, patients use various social networks such as Twitter, blog, and various forums to provide a review of medicines and other pharmaceutical information. As the twitter is used by millions of users, there are high possibilities that people are using twitter more often than any other reporting tool. Thus, we claim that social media such as Twitter is going to be a very good
Data Loading...
