Journal of Medical Internet Research

View query in Help articles search

Field

e.g. mhealth

Operator

Field

Criteria

e.g. John Smith or 0000-0001-6479-5330

Exact Match for Criteria

Yes No

Search Results (1 to 8 of 8 Results)

Download search results: CSV END BibTex RIS

Sort by:

News Media Framing of Suicide Circumstances and Gender: Mixed Methods Analysis

We generated a unique vector representation for each news article in our data set to answer research question 1 and to analyze how suicide is framed using the SOSS dimensions. We again used BERT, programmed in Python (Python Software Foundation) using Hugging Face transformers [31], to extract a sentence-level embedding representation for each news article. This allowed us to capture nuanced attributes such as context, framing and writing style, and the relationship between words in a sentence.

Jasmine C Foriest, Shravika Mittal, Eugenia Kim, Andrea Carmichael, Natalie Lennon, Steven A Sumner, Munmun De Choudhury

JMIR Ment Health 2024;11:e49879

Download Citation: END BibTex RIS

Download PDF Download XML

Emerging Trends of Self-Harm Using Sodium Nitrite in an Online Suicide Community: Observational Study Using Natural Language Processing Analysis

Since 2019, a growing trend of using sodium nitrite (SN), a common food additive, for self-harm has been reported [6,7]. SN has traditionally been used as a food preservative and coloring agent, in addition to use as a corrosion inhibitor. As such, it is widely available for purchase. An alarming development has been the sale of “suicide kits” in online marketplaces, which comprise SN in addition to instructional material on attempting suicide [8].

Sudeshna Das, Drew Walker, Swati Rajwal, Sahithi Lakamana, Steven A Sumner, Karin A Mack, Wojciech Kaczkowski, Abeed Sarker

JMIR Ment Health 2024;11:e53730

Download Citation: END BibTex RIS

Download PDF Download XML

Using Transformer-Based Topic Modeling to Examine Discussions of Delta-8 Tetrahydrocannabinol: Content Analysis

BERTopic uses a probabilistic generative topic model, which uses deep learning models to determine the topics in a collection of documents. BERTopic first represents each post as an embedding, leveraging a large pretrained language model. As a next step, the dimensionality of the embeddings is reduced using uniform manifold approximation and projection, an algorithm that outperforms traditional techniques such as principal component analysis in the setting of high-dimensional data [23].

Brandi Patrice Smith, Brooke Hoots, Lara DePadilla, Douglas R Roehler, Kristin M Holland, Daniel A Bowen, Steven A Sumner

J Med Internet Res 2023;25:e49469

Download Citation: END BibTex RIS

Download PDF Download XML

Web-Based Social Networks of Individuals With Adverse Childhood Experiences: Quantitative Study

This work involved a multistep pipeline to first create a classifier to identify ACE content from social media posts, deploy the classifier to a large body of social media posts, and compute network statistics from such information. In Textbox 1, we provide a short description of the terms introduced in the following subsections. Accuracy This is a measure of classification performance, defined by the fraction of the correct prediction.

Yiding Cao, Suraj Rajendran, Prathic Sundararajan, Royal Law, Sarah Bacon, Steven A Sumner, Naoki Masuda

J Med Internet Res 2023;25:e45171

Download Citation: END BibTex RIS

Download PDF Download XML

Characterizing and Identifying the Prevalence of Web-Based Misinformation Relating to Medication for Opioid Use Disorder: Machine Learning Approach

Health misinformation is defined as a health-related claim of a fact that is currently false because of a lack of scientific evidence [22]. In general, misinformation is usually attributed to misconceptions and is not intended to cause harm. Disinformation is false information that is created deliberately to cause harm, with motivations that are often social, political, or financial.

Mai ElSherief, Steven A Sumner, Christopher M Jones, Royal K Law, Akadia Kacha-Ochana, Lyna Shieber, LeShaundra Cordier, Kelly Holton, Munmun De Choudhury

J Med Internet Res 2021;23(12):e30753

Download Citation: END BibTex RIS

Download PDF Download XML

Suicide Risk and Protective Factors in Online Support Forum Posts: Annotation Scheme Development and Validation Study

Current quantifiable risk evaluations of suicidality map to a single perspective of evaluating risk, which focuses on aggregated notions of riskiness that may determine a response from a clinician. Instead of collapsing the notion of risk into a singular point, clinical and public health professionals instead often examine and track risk factors or attributes and characteristics that increase an individual’s likelihood of attempting suicide in the future [18].

Stevie Chancellor, Steven A Sumner, Corinne David-Ferdon, Tahirah Ahmad, Munmun De Choudhury

JMIR Ment Health 2021;8(11):e24471

Download Citation: END BibTex RIS

Download PDF Download XML