Real time predictive analytics and actionable insights for water intake on wearable device

How, Chun Siong (2017) Real time predictive analytics and actionable insights for water intake on wearable device. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

	Text (24 Pages) Real time predictive analytics and actionable insights for water intake on wearable device.pdf - Submitted Version Download (1MB)
	Text (Full Text) Real time predictive analytics and actionable insights for water intake on wearable device.pdf - Submitted Version Restricted to Registered users only Download (3MB)

Abstract

Now a day, many people will try to use wearable device to track their fitness and heath performance. In the market, there are many brands of the fitness band but majority of them do not provide any intelligence features inside like actionable insight or suggestion to user especially in suggestion for water requirement. Water requirement is based on respiratory water loss, urine loss, insensible water loss, metabolic water production and fecal loss. The total water requirement is sum of respiratory water loss, urine loss, insensible water loss and fecal loss, then minus metabolic water production. The wearable device will send the real time data into database and it will analyze to give suitable actionable insight like cup of water intake and dehydration level to user. There are three objectives of this project which are to develop a real time module that can give predictive analytics for volume of water requirement, to generate actionable insight from predictive analytics, and to let user monitoring the water intake requirement for whole day. The technologies that used to develop system are python and hypertext markup language(HTML). The Fuzzy model will be used in predicting water requirement. First stage is to collect user detail data like weight, height, gender and age to generate fuzzy membership functions of total water requirement. Second stage is collect the data required for model like heart rate, surrounding temperature and activity status from wearable device. After that, input the data into fuzzy model that can produce water requirement per second. Then, this system able to give suitable actionable insight to user based on previous analytic. The user detail data will be used in updating model automatically once user updates detail to system. The test strategies that use to test the system after implementation is black box testing. As a result, fuzzy model is performed well in thus system and the objectives are fulfilled. Besides that, the system contribute to the industri area like wearable device company. For further improvement, more research on AI technique has to be made so that the suitable technique is used for this system.

Actions (login required)

View Item