%0 Journal Article %I JMIR Publications %V 5 %N %P e53365 %T Development of Depression Data Sets and a Language Model for Depression Detection: Mixed Methods Study %A Tumaliuan,Faye Beatriz %A Grepo,Lorelie %A Jalao,Eugene Rex %+ Department of Industrial Engineering and Operations Research, University of the Philippines Diliman, Melchor Hall, Magsaysay Avenue, Quezon City, 1101, Philippines, 63 9176593613, fayetumaliuan@gmail.com %K depression data set %K depression detection %K social media %K natural language processing %K Filipino %D 2024 %7 4.9.2024 %9 Original Paper %J JMIR Data %G English %X Background: Depression detection in social media has gained attention in recent years with the help of natural language processing (NLP) techniques. Because of the low-resource standing of Filipino depression data, valid data sets need to be created to aid various machine learning techniques in depression detection classification tasks. Objective: The primary objective is to build a depression corpus of Philippine Twitter users who were clinically diagnosed with depression by mental health professionals and develop from this a corpus of depression symptoms that can later serve as a baseline for predicting depression symptoms in the Filipino and English languages. Methods: The proposed process included the implementation of clinical screening methods with the help of clinical psychologists in the recruitment of study participants who were young adults aged 18 to 30 years. A total of 72 participants were assessed by clinical psychologists and provided their Twitter data: 60 with depression and 12 with no depression. Six participants provided 2 Twitter accounts each, making 78 Twitter accounts. A data set was developed consisting of depression symptom–annotated tweets with 13 depression categories. These were created through manual annotation in a process constructed, guided, and validated by clinical psychologists. Results: Three annotators completed the process for approximately 79,614 tweets, resulting in a substantial interannotator agreement score of 0.735 using Fleiss κ and a 95.59% psychologist validation score. A word2vec language model was developed using Filipino and English data sets to create a 300-feature word embedding that can be used in various machine learning techniques for NLP. Conclusions: This study contributes to depression research by constructing depression data sets from social media to aid NLP in the Philippine setting. These 2 validated data sets can be significant in user detection or tweet-level detection of depression in young adults in further studies. %R 10.2196/53365 %U https://data.jmir.org/2024/1/e53365 %U https://doi.org/10.2196/53365 %0 Journal Article %I JMIR Publications %V 1 %N 1 %P e22436 %T A Comparison of Blood Pressure Data Obtained From Wearable, Ambulatory, and Home Blood Pressure Monitoring Devices: Prospective Validation Study %A Islam,Sheikh Mohammed Shariful %A Maddison,Ralph %+ Institute for Physical Activity and Nutrition, Deakin University, 221 Burwood Highway, Burwood, Melbourne, 3125, Australia, 61 451733373, shariful@deakin.edu.au %K wearable devices %K mobile phones %K blood pressure %K ambulatory blood pressure monitoring %K home blood pressure devices %D 2020 %7 4.11.2020 %9 Original Paper %J JMIR Data %G English %X Background: Blood pressure (BP) is an important marker for cardiovascular health. However, a person’s BP data cannot usually be obtained simultaneously from different sources. Objective: This study aimed to analyze and compare BP data obtained from 3 different sources, namely, wearable, ambulatory, and home BP monitoring devices. Methods: During recruitment, we recorded participants’ BP using a standardized digital BP monitoring device and simultaneously over 24 hours using wearable and ambulatory devices. In addition, participants’ BP was measured over 7 days using wearable and home BP monitoring devices. Data from the wearable BP monitoring devices were extracted. The 24-hour ambulatory BP data were downloaded from the device to a computer. Home BPs were recorded 3 times per day (in the morning, afternoon, and evening, at regular times convenient to the participants) for 7 days and on a BP sheet. Results: A total of 9090 BP measurements were collected from 20 healthy volunteer participants (females: n=10; males: n=10, mean age 20.3 years, SD 5.4 years). The mean (SD) systolic BP and diastolic BP values measured at enrollment were 112.35 (9.79) mm Hg and 73.75 (9.14) mm Hg, respectively. The 24-hour mean (SD) systolic BP and diastolic BP values measured using the wearable device were 125 (5) mm Hg and 77 (9) mm Hg, respectively. The 24-hour mean (SD) systolic BP and diastolic BP values recorded using the ambulatory device were 126 (10) mm Hg and 75 (6) mm Hg, respectively. The 7-day mean (SD) systolic BP and diastolic BP values measured using the wearable device were 125 (4) mm Hg and 77 (3) mm Hg, respectively. The 7-day mean (SD) systolic BP and diastolic BP values measured using the home device were 112 (10) mm Hg and 71 (8) mm Hg, respectively. Conclusions: Our datasets serve as the basis for further studies where these data can be combined reasonably with data from similar studies to understand the impact of different devices on BP measurement. Moreover, the BP data acquired noninvasively from wearable, ambulatory, and home devices can be integrated with similar data from other studies to determine the utility of wearable BP monitoring devices in different groups of people. %R 10.2196/22436 %U https://data.jmir.org/2020/1/e22436 %U https://doi.org/10.2196/22436