Proximity-based active learning on streaming data for eating moment recognition using wearable sensors

Marjan Nourollahi Darabad

Back

Proximity-based active learning on streaming data for eating moment recognition using wearable sensors

Thesis

Open access

Proximity-based active learning on streaming data for eating moment recognition using wearable sensors

Marjan Nourollahi Darabad

Washington State University

Master of Science (MS), Washington State University

2019

Handle:

https://hdl.handle.net/2376/103609

Files and links (1)

pdf

M_NourollahiDarabad_080819694.52 kBDownload View

Open Access

Abstract

Machine learning.;Streaming technology (Telecommunications);Diet.;Wearable technology.

Detecting when eating occurs is an essential step toward automatic dietary monitoring, medication adherence assessment, and diet-related health interventions. Wearable technologies play a central role in designing unubtrusive diet monitoring solutions by leveraging machine learning algorithms that work on time-series sensor data to detect eating moments. While much research has been done on developing activity recognition and eating moment detection algorithms, the performance of the detection algorithms drops substantialy when the model trained with one user is utilized by a new user. To facilitate development of personalized models, PALS, ProximityBased Active Learning on Streaming Data, a novel proximity-based model is proposed for recognizing eating gestures with the goal of significantly decreasing the need for labeled data with new users. Particularly, an optimization problem is proposed to perform active learning under limited query budget by leveraging unlabeled data. The extensive data analysis in this thesis on data collected in both controlled and uncontrolled settings indicates that PALS achieves 22% higher F1-score for detecting eating events compared to the existing approaches.

Metrics

7 File views/ downloads

16 Record Views

Details

Title: Proximity-based active learning on streaming data for eating moment recognition using wearable sensors
Creators: Marjan Nourollahi Darabad
Contributors: Hassan Ghasemzadeh (Degree Supervisor)
Awarding Institution: Washington State University
Academic Unit: Electrical Engineering and Computer Science, School of
Theses and Dissertations: Master of Science (MS), Washington State University
Publisher: Washington State University; [Pullman, Washington] :
Identifiers: 99900525034001842
Language: English
Resource Type: Thesis