Blog

For STEM admissions, achieving high grades and participating in extracurricular activities, such as research paper publications, are becoming increasingly vital. Engaging in research involves high school students participating directly in university-level research to showcase their interest in specific departments and demonstrate their unique contributions. 

It's essential to select topics aligned with their strongest STEM subjects, aiming for practical, real-world applications. Exploring emerging scientific technologies can be an effective way to choose compelling research topics. Since research is self-directed, it's important to choose topics that allow them to articulate how their innovative ideas can contribute to future societal progress. Topics can be proposed based on internet articles and journals.


 

1. Engineering & Programming

※ Reference Article: https://astroblog.cosmobc.com/internet-change-space-exploration/

Topic

Proposing a solution to mow grass without interfering with telescopes based on IoT's impact on space exploration
 

Issues

- Businesses struggle to address security concerns arising from the vast data generated by IoT.

- Astronomers have raised concerns about connectivity issues with IoT devices.

- Legal constraints exist regarding the use of IoT technology.

- NRAO astronomers warn that iRobot lawnmowers can disrupt multiple telescopes.
 

Hypothesis

Creating a "Quiet Zone" around observatories, where wireless signals are blocked, allows iRobot lawnmowers to utilize IoT for grass mowing without affecting telescopes.
 

Research Methodology

Implementing machine learning, iRobot lawnmowers will analyze data on electricity consumption, grass height, weather conditions, and mowing paths to optimize efficiency and timing. Demonstrate that the "Quiet Zone" effectively prevents iRobot lawnmowers from interfering with telescopes. Design an AI program enabling iRobot lawnmowers to operate without disrupting telescopes while mowing grass.

2. Machine Learning

※ Reference Article: https://blog.naver.com/gdpresent/223033752466 

Topic

Machine Learning for Panel Data Analysis
 

Hypothesis

Utilizing decision trees in machine learning can significantly improve prediction accuracy for panel data.
 

Research Methodology

Based on the Boston Housing dataset, employ various techniques including linear regression, decision trees, dendrological analysis, addressing the bias-variance tradeoff, hyperparameter testing, and methodologies such as training/test/validation, bagging, random forests, boosting, and support vector machines to validate enhancements over traditional linear regression.

 

3. Social Science

※ Reference Article: https://www.fda.gov/medical-devices/digital-health-center-excellence/what-digital-health 

Topic

The Impact of Telemedicine Features in Digital Health Management Applications on Healthcare and Society
 

Hypothesis

Introducing telemedicine features in digital health management applications may lead to healthcare workforce substitution, potentially impacting public health negatively.
 

Research Methodology

Analyze the practical effects of telemedicine features, conduct probabilistic surveys to determine preferences between telemedicine and pharmacy prescriptions based on age groups and medical conditions. Investigate the rate of healthcare workforce substitution and assess the risk associated with prescriptions under telemedicine. Compare average health outcomes relative to the rate of workforce substitution.