Analyzing Sensor-Based Individual and Population Behavior Patterns via Inverse Reinforcement Learning

Lin, Beiyu; Cook, Diane J.

doi:10.3390/s20185207

Cited by 4 publications

(2 citation statements)

References 44 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Viterbi algorithm was used to test the validity of the obtained model over a set of testing days. The approach presented in [28] learns the behavioral routine of an individual using an inverse reinforcement learningbased solution and by considering the Spatio-temporal information. For this, the sequential decision-making space of the resident in a smart home is modeled as a Markov Decision Process, and then it is used to learn the resident's behavioral routine via relative entropy inverse reinforcement learning.…”

Section: Related Workmentioning

confidence: 99%

Deep Learning, Mining, and Collaborative Clustering to Identify Flexible Daily Activities Patterns

Chifu

Pop

Rancea

et al. 2022

Sensors

View full text Add to dashboard Cite

The monitoring of the daily life activities routine is beneficial, especially in old age. It can provide relevant information on the person’s health state and wellbeing and can help identify deviations that signal care deterioration or incidents that require intervention. Existing approaches consider the daily routine as a rather strict sequence of activities which is not usually the case. In this paper, we propose a solution to identify flexible daily routines of older adults considering variations related to the order of activities and activities timespan. It combines the Gap-BIDE algorithm with a collaborative clustering technique. The Gap-BIDE algorithm is used to identify the most common patterns of behavior considering the elements of variations in activities sequence and the period of the day (i.e., night, morning, afternoon, and evening) for increased pattern mining flexibility. K-means and Hierarchical Clustering Agglomerative algorithms are collaboratively used to address the time-related elements of variability in daily routine like activities timespan vectors. A prototype was developed to monitor and detect the daily living activities based on smartwatch data using a deep learning architecture and the InceptionTime model, for which the highest accuracy was obtained. The results obtained are showing that the proposed solution can successfully identify the routines considering the aspects of flexibility such as activity sequences, optional and compulsory activities, timespan, and start and end time. The best results were obtained for the collaborative clustering solution that considers flexibility aspects in routine identification, providing coverage of monitored data of 89.63%.

show abstract

Section: Related Workmentioning

confidence: 99%

Deep Learning, Mining, and Collaborative Clustering to Identify Flexible Daily Activities Patterns

Chifu

Pop

Rancea

et al. 2022

Sensors

View full text Add to dashboard Cite

show abstract

“…In order to analyze the data from sensor networks and individual/group behavior, an understanding of social learning theory is required because behavioral routines can be found in the sensor data. (13) Game theory is used to assay the outcomes or consequences of each choice in an entire decision-making procedure. Game theory was proposed to prove the concept of "taking the minimum in a negative situation" and to construct the "zero-sum game".…”

Section: Introductionmentioning

confidence: 99%

Important Elements of Sensor Technology and Data Management and Related Education

Chan¹,

Lai²,

Hsieh³

et al. 2023

Sensors and Materials

View full text Add to dashboard Cite

The demand for both sensors and a dedicated workforce has been increasing rapidly; an appropriate curriculum is needed to train this workforce. This study was conducted to identify elements that need to be included in the curriculum to teach sensor technology and the associated data management effectively. In this study, interviews with experts and a survey by questionnaire were conducted. The data were analyzed with an analytical network process (ANP), and statistical analysis including ANOVA, regression analysis, and factor analysis. As a result, four criteria (problem analysis, pattern recognition, abstraction of problems, and finding solutions) were defined with 12 subcriteria including faculty education in the curriculum, self-directed learning in the curriculum, committees and working groups to structure the curriculum, planning, organizing and managing the curriculum, learning outcomes, feedback from students and teaching staff, advice of experts, training assistants, appropriate assessment, appropriate education, appropriate evaluation, and recruiting students. Analytical skills, problem-solving, and decision-making were found to be alternatives (elements in education) in the analysis. The results of ANP and other statical analyses indicated that among the criteria, problem analysis is more important than the others in education for sensor technology and data management. Among alternatives (educational elements), analytical skills are more important than problem-solving and decision-making. Therefore, education in sensor-related subjects needs to be more focused on analyzing and detecting problems in sensor data. These results provide the basis for creating a curriculum for education in sensor-related technologies.

show abstract

Artificial Intelligence for Cognitive Health Assessment: State-of-the-Art, Open Challenges and Future Directions

et al. 2023

View full text Add to dashboard Cite

The subjectivity and inaccuracy of in-clinic Cognitive Health Assessments (CHA) have led many researchers to explore ways to automate the process to make it more objective and to facilitate the needs of the healthcare industry. Artificial Intelligence (AI) and machine learning (ML) have emerged as the most promising approaches to automate the CHA process. In this paper, we explore the background of CHA and delve into the extensive research recently undertaken in this domain to provide a comprehensive survey of the state-of-the-art. In particular, a careful selection of significant works published in the literature is reviewed to elaborate a range of enabling technologies and AI/ML techniques used for CHA, including conventional supervised and unsupervised machine learning, deep learning, reinforcement learning, natural language processing, and image processing techniques. Furthermore, we provide an overview of various means of data acquisition and the benchmark datasets. Finally, we discuss open issues and challenges in using AI and ML for CHA along with some possible solutions. In summary, this paper presents CHA tools, lists various data acquisition methods for CHA, provides technological advancements, presents the usage of AI for CHA, and open issues, challenges in the CHA domain. We hope this first-of-its-kind survey paper will significantly contribute to identifying research gaps in the complex and rapidly evolving interdisciplinary mental health field.

show abstract

Analyzing Sensor-Based Individual and Population Behavior Patterns via Inverse Reinforcement Learning

Cited by 4 publications

References 44 publications

Deep Learning, Mining, and Collaborative Clustering to Identify Flexible Daily Activities Patterns

Deep Learning, Mining, and Collaborative Clustering to Identify Flexible Daily Activities Patterns

Important Elements of Sensor Technology and Data Management and Related Education

Artificial Intelligence for Cognitive Health Assessment: State-of-the-Art, Open Challenges and Future Directions

Contact Info

Product

Resources

About