Search Articles

View query in Help articles search

Search Results (1 to 10 of 10 Results)

Download search results: CSV END BibTex RIS


Drug Abuse Ontology to Harness Web-Based Data for Substance Use Epidemiology Research: Ontology Development Study

Drug Abuse Ontology to Harness Web-Based Data for Substance Use Epidemiology Research: Ontology Development Study

We evaluated our approach using the average F1-score for each user type individual (P), informed agency (I), and retailer (R). The F1 scores for the individual classes P, I, and R were 95%, 42%, and 73%, respectively. The descriptive statistics of the training set at the Twitter user account level used for this study, which involved semantic filtering [84] using the DAO, are shown in Table 2. Architecture of the e Drug Trends project.

Usha Lokala, Francois Lamy, Raminta Daniulaityte, Manas Gaur, Amelie Gyrard, Krishnaprasad Thirunarayan, Ugur Kursuncu, Amit Sheth

JMIR Public Health Surveill 2022;8(12):e24938

Knowledge-Infused Abstractive Summarization of Clinical Diagnostic Interviews: Framework Development Study

Knowledge-Infused Abstractive Summarization of Clinical Diagnostic Interviews: Framework Development Study

Sheth et al [6,8] define knowledge-infused learning as “the exploitation of domain knowledge and application semantics to enhance existing artificial intelligence methods by infusing relevant conceptual information into a statistical and data-driven computational approach,” which in this study is integer linear programming (ILP).

Gaur Manas, Vamsi Aribandi, Ugur Kursuncu, Amanuel Alambo, Valerie L Shalin, Krishnaprasad Thirunarayan, Jonathan Beich, Meera Narasimhan, Amit Sheth

JMIR Ment Health 2021;8(5):e20865

Identifying Key Topics Bearing Negative Sentiment on Twitter: Insights Concerning the 2015-2016 Zika Epidemic

Identifying Key Topics Bearing Negative Sentiment on Twitter: Insights Concerning the 2015-2016 Zika Epidemic

Before analysis, tweets were preprocessed by removing non-American Standard Code for Information Interchange (ASCII) characters, capital letters, retweet indicators, numbers, screen handles (@username), punctuation, URLs, whitespaces, single characters such as p that do not convey any meaning about topics in the corpus, and stop words such as and, so, etc. A random sample of 1467 tweets was annotated as relevant versus not by 3 microbiology and immunology experts and used as the relevancy ground truth.

Ravali Mamidi, Michele Miller, Tanvi Banerjee, William Romine, Amit Sheth

JMIR Public Health Surveill 2019;5(2):e11036

“How Is My Child’s Asthma?” Digital Phenotype and Actionable Insights for Pediatric Asthma

“How Is My Child’s Asthma?” Digital Phenotype and Actionable Insights for Pediatric Asthma

Using the NHLBI guidelines given in Table 3 [20], we have developed thresholds for DPS (DPS-P and DPS-T) as shown in Table 4 (eg, if the DPS≥1, then the child’s asthma is very poorly controlled). The thresholds for the 3 control levels have been chosen to make the DPS-P and ACT scores comparable when they are available over the same period.

Utkarshani Jaimini, Krishnaprasad Thirunarayan, Maninder Kalra, Revathy Venkataraman, Dipesh Kadariya, Amit Sheth

JMIR Pediatr Parent 2018;1(2):e11988

“When ‘Bad’ is ‘Good’”: Identifying Personal Communication and Sentiment in Drug-Related Tweets

“When ‘Bad’ is ‘Good’”: Identifying Personal Communication and Sentiment in Drug-Related Tweets

One-tailed t test statistic was used to determine which classifiers performed significantly better (P Source classification (Approach 1) that used short URLs demonstrated good performance (Table 1 A). SVM algorithm applied to multiclass classification task achieved a macro average F-score of 0.7972, which was not significantly higher compared with LR (P=.09) or NB (P=.27) performance (Table 1 A).

Raminta Daniulaityte, Lu Chen, Francois R Lamy, Robert G Carlson, Krishnaprasad Thirunarayan, Amit Sheth

JMIR Public Health Surveill 2016;2(2):e162