What would be the best big data analytical tools to determine changes in behavioral health (e.g., addiction behaviors)? The data used to predict behavioral health changes are secondary data which consists of environmental data (e.g., air, water quality, climate data), socioeconomic data (e.g., income, jobs, social status), and social media activities. I having a feeling that some sort of optimization tool should be used to conduct the analysis for this type of research questions. Any suggestions will be highly appreciated :)
Dear Tyson Ang,
Yes, I agree that the data used to predict changes in health behavior are secondary data, consisting of environmental data (e.g. data on air condition, water quality, climate change, climate zone), socio-economic determinants (e.g. income, work, social status) and activity in various media, mainly in new online media, including social media. This is how sentiment analysis is used for this type of research processes, carried out on the basis of large sets of data collected from selected websites and also other sources of information concerning, on the one hand, specific persons or entities and related to cross-sectional, sectoral data, etc. These types of analytical processes are carried out on the basis of Big Data Analytics platforms. If the sentiment analysis is carried out cyclically on constantly updated data sets concerning the same objects, then the results of these analyzes may be an additional source of data for the purposes of forecasting future continuation and / or changes in specific development trends. Best wishes,
Dariusz Prokopowicz
Tyson Ang Here are some articles I can recommend that will help your approach to your work.
Article Fostering Engagement in Health Behavior Change: Iterative De...
Article Youth Risk Behavior Surveillance - United States, 2017
Article Using Mobile Apps to Promote a Healthy Lifestyle Among Adole...
Kindly check https://www.researchgate.net/publication/332070483_Big_Data_Analytics_and_AI_in_Mental_Healthcare
Also check https://www.researchgate.net/publication/325163585_Emerging_Technology_and_Data_Analytics_for_Behavioral_Health_Preprint
Big data can address some of the areas mentioned.
One of the cautions in clinical predictions is the distinction between identifying a probability rate for a group vs. using the analytic findings at the individual case level. For example one can reliably predict that a percentage of a certain group will have a specified bimodal outcome (e.g., recovering vs. not recovering; survive vs. died; etc.), but cannot predict which individuals fall into which outcome. If you can predict that 90% of a specified group will have a positive outcome, you will still be wrong for one in ten cases if you predict a positive outcome.
Akpofure A Enughwure Chinaza Godswill Awuchi Thank you so much for the references
Norman G Hoffmann I agree with your cautions. What would be your recommendation?
I am thinking some sort of optimization technique should be used to understand environmental factors that contribute to risky/non-risky behaviors.
Optimization techniques can be used to guide strategies and develop knowledge for achieving efficiency and efficacy. However, when dealing with individuals, there is no substitute for sound clinical judgment from experienced clinicians. That said, AI can be very accurate for diagnostics and treatment planning. Paul Meehl in the 1950s indicated that empirical practices could outperform clinical judgment. It is in the area of prediction that no approach is 100% accurate.
- Badges
- Science topic
- Similar topics
- Social Science
- Media
- Social Media