The Application of Text Mining Algorithms In Summarizing Trends in Anti-Epileptic Drug Research

Abstract

Content summarization is an important area of research in traditional data mining. The volume of studies published on anti-epileptic drugs (AED) has increased exponentially over the last two decades, making it an important area for the application of text mining based summarization algorithms. In the current study, we use text analytics algorithms to mine and summarize 10,000 PubMed abstracts related to anti-epileptic drugs published within the last 10 years. A Text Frequency -Inverse Document Frequency based filtering was applied to identify drugs with highest frequency of mentions within these abstracts. The US Food and Drug database was scrapped and linked to the results to quantify the most frequently mentioned modes of action and elucidate the pharmaceutical entities marketing these drugs. A sentiment analysis model was created to score the abstracts for sentiment positivity or negativity. Finally, a modified Latent Dirichlet Allocation topic model was generated to extract key topics associated with the most frequently mentioned AEDs. We found the top five most common drugs that appeared from the analysis were Gabapentin, Levetiracetam, Topiramate, Lamotrigine and Acetazolamide. We further listed the key topics associated with these drugs and the overall positive or negative sentiment associated with them. Results of this study provide accurate and data intensive insights on the progress of anti-epileptic drug research.In this study, we analyze 10,000 PubMed abstracts related to AEDs with the end goal of content summarization and Due to the coupling between θ and φ in the integrand (above equation), the precise implication in mLDA is obstinate (Blei et al., 2003).The number of topics was selected according to the Rate of Perplexity Change (RPC) previously described by Zhao and colleagues (Zhao et al., 2015). This algorithm yielded two key topics on average which were further curated manually.

Pediatric literature trends: high-level analysis using text-mining

“…However, the number of pediatric publications increased exponentially over the past decades making it impossible to manually summarize a topic. 4 In text-mining, information is extracted from texts using computer algorithms. 5 Text-mining can be applied to identify trends and to investigate the dynamics in a research field.…”

“…5 Text-mining can be applied to identify trends and to investigate the dynamics in a research field. 4,[6][7][8] Examples include anti-epileptic drug research, 9 adolescent substance abuse, 10 and Diatom research. 9 The application of such algorithms has not been frequently used in pediatric research.…”

Unsupervised Machine Learning for Co/Multimorbidity Analysis

“…The R packages "Cluster" (v 2.0.7.1) and "Dendextend' (1.8.0) was employed for this analysis. Significance testing for normal, non-normal and binary data was performed as described previously by the authors in other studies (Fleming et al, 2018;Hester et al, 2016;Karkare et al, 2014;Singh, 2015Singh, , 2016Singh et al, 2016;Singh, He, McNamara, & Danzer, 2013;Singh, Karkare, Baswan, & Singh, 2018;Singh, LaSarge, An, McAuliffe, & Danzer, 2015;Singh, Singh, Fatima, Kubo, & Singh, 2008;Singh & Karkare, 2017;Singh & Singh, 2017). De-identified data and the statistical analysis R code used was uploaded to an online repository.…”