We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Blockchain associated machine learning and IoT based hypoglycemia detection system with auto-injection feature

Abstract: Hypoglycemia is an unpleasant phenomenon caused by low blood glucose. The disease can lead a person to death or a high level of body damage. To avoid significant damage, patients need sugar. The research aims at implementing an automatic system to detect hypoglycemia and perform automatic sugar injections to save a life. Receiving the benefits of the internet of things (IoT), the sensor data was transferred using the hypertext transfer protocol (HTTP) protocol. To ensure the safety of health-related data, blockchain technology was utilized. The glucose sensor and smartwatch data were processed via Fog and sent to the cloud. A Random Forest algorithm was proposed and utilized to decide hypoglycemic events. When the hypoglycemic event was detected, the system sent a notification to the mobile application and auto-injection device to push the condensed sugar into the victims body. XGBoost, k-nearest neighbors (KNN), support vector machine (SVM), and decision tree were implemented to compare the proposed models performance. The random forest performed 0.942 testing accuracy, better than other models in detecting hypoglycemic events. The systems performance was measured in several conditions, and satisfactory results were achieved. The system can benefit hypoglycemia patients to survive this disease.
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computers and Society (cs.CY); Signal Processing (eess.SP)
Journal reference: Indonesian Journal of Electrical Engineering and Computer Science, Vol. 27, No. 1, pp. 447-455, July 2022
DOI: 10.11591/ijeecs.v27.i1.pp447-455
Cite as: arXiv:2208.02222 [cs.LG]
  (or arXiv:2208.02222v1 [cs.LG] for this version)

Submission history

From: Al-Akhir Nayan [view email]
[v1] Wed, 27 Jul 2022 03:11:41 GMT (621kb)

Link back to: arXiv, form interface, contact.