Unsupervised labelling of sequential data for location identification in indoor environments

Perez, Iker; Pinchin, James; Brown, Michael; Blum, Jesse; Sharples, Sarah

doi:10.1016/j.eswa.2016.06.003

Cited by 5 publications

(3 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Indoor positioning with Wi-Fi faces the major problem of signal attenuations [39], which is a common cause of faulty measurements. Apart from signal processing works [40], numerous Wi-Fi-based studies have examined indoor positioning approaches [31], reporting either accuracy improvement [8,41,42] or applying machine learning techniques, such as decision trees [23], unsupervised labelling on sequential data [43], unsupervised clustering for multi-floor environments [9] and online sequential extreme learning [44,45]. Bayesian models have also been utilized eliminating the problem of training data.…”

Section: Related Workmentioning

confidence: 99%

An Ensemble Filter for Indoor Positioning in a Retail Store Using Bluetooth Low Energy Beacons

Stavrou

Bardaki

Papakyriakopoulos

et al. 2019

Sensors

View full text Add to dashboard Cite

This paper has developed and deployed a Bluetooth Low Energy (BLE) beacon-based indoor positioning system in a two-floor retail store. The ultimate purpose of this study was to compare the different indoor positioning techniques towards achieving efficient position determination of moving customers in the retail store. The innovation of this research lies in its context (the retail store) and the fact that this is not a laboratory, controlled experiment. Retail stores are challenging environments with multiple sources of noise (e.g., shoppers’ moving) that impede indoor localization. To the best of the authors’ knowledge, this is the first work concerning indoor localization of consumers in a real retail store. This study proposes an ensemble filter with lower absolute mean and root mean squared errors than the random forest. Moreover, the localization error is approximately 2 m, while for the random forest, it is 2.5 m. In retail environments, even a 0.5 m deviation is significant because consumers may be positioned in front of different store shelves and, thus, different product categories. The more accurate the consumer localization, the more accurate and rich insights on the customers’ shopping behavior. Consequently, retailers can offer more effective customer location-based services (e.g., personalized offers) and, overall, better consumer localization can improve decision making in retailing.

show abstract

Section: Related Workmentioning

confidence: 99%

An Ensemble Filter for Indoor Positioning in a Retail Store Using Bluetooth Low Energy Beacons

Stavrou

Bardaki

Papakyriakopoulos

et al. 2019

Sensors

View full text Add to dashboard Cite

show abstract

“…Secondary healthcare systems around the world are under increasing pressure ( [1,2] and specific working groups ( [20]); examples of such work include [21] and [22]. In this paper, we explore the use of electronic task-management for the study of OoH workload in secondary facilities.…”

Section: Introductionmentioning

confidence: 99%

Out of hours workload management: Bayesian inference for decision support in secondary care

Perez

Brown

Pinchin

et al. 2016

Artificial Intelligence in Medicine

Self Cite

View full text Add to dashboard Cite

Objective: In this paper, we aim to evaluate the use of electronic technologies in Out of Hours (OoH) task-management for assisting the design of effective support systems in health care; targeting local facilities, wards or specific working groups. In addition, we seek to draw and validate conclusions with relevance to a frequently revised service, subject to increasing pressures.Methods and Material : We have analysed 4 years of digitised demand-data extracted from a recently deployed electronic task-management system, within the Hospital at Night setting in two jointly coordinated hospitals in the United Kingdom. The methodology employed relies on Bayesian inference methods and parameter-driven state-space models for multivariate series of count data.Results: Main results support claims relating to (i) the importance of datadriven staffing alternatives and (ii) demand forecasts serving as a basis to intelligent scheduling within working groups. We have displayed a split in workload patterns across groups of medical and surgical specialities, and sustained assertions regarding staff behaviour and work-need changes according to shifts or days of the week. Also, we have provided evidence regarding the relevance of day-to-day planning and prioritisation.Conclusions: The work exhibits potential contributions of electronic tasking alternatives for the purpose of data-driven support systems design; for scheduling, prioritisation and management of care delivery. Response to reviewersGeneral Comments: We are grateful to all reviewers for taking the time to review our paper and make a number of helpful comments that we believe have significantly improved the quality of the manuscript. All comments have been, according to their nature, merged, described and addressed here: 5Comment 1: The major concern shared by reviewers relates to a lack practical relevance of the results presented; for instance, managerial conclusions are not well exploited, the paper lacks a managerial perspective of the output from professionals involved in the study, or results do not extrapolate to different settings neither is it clear why they support the decision making. 10Response to Comment 1: The authors agree that the previous version of the manuscript offered a mostly descriptive study of the data and lacked relevant interpretations and conclusions drawn from results. In addition, we believe we were not successful in clearly conveying the aims and scope of our work; yet, we have made the efforts to undertake the alterations necessary in order to correct 15 these problems.Nationwide hospital settings are indeed subject to disparate workload pressures; these may drastically vary within distant geographical regions and in relation to specialities covered by each facility. The focus of our research team and this work relates to support systems for effective management that can 20 target local facilities and working groups. This is indeed a great problem, in that policy alterations and intelligent design of support systems (for instan...

show abstract

“…Regarding accuracy improvement, Torres-Sospedra et al (2015) focus on the optimal distance function and conclude to the selection of the best configuration for an indoor positioning algorithm. Machine learning techniques have also been used for indoor positioning, such as decision trees (Yim, 2008), unsupervised labelling on sequential data (Perez et al, 2016), unsupervised clustering for multi-floor environments (Campos et al, 2014) and Online Sequential Extreme Learning (Zou et al, 2014;.…”

Section: Indoor Positioning Wireless Infrastructure Technologiesmentioning

confidence: 99%

Indoor localization systems for retail stores

Σταύρου¹

View full text Add to dashboard Cite

Η συμπεριφορά του καταναλωτή μέσα στο κατάστημα αποτελεί ερευνητική περιοχή ενδιαφέροντος για περισσότερα από 60 χρόνια με στόχο να εντοπίσει διάφορα μοτίβα που μπορούν να προσφέρουν αξία. Η κατανόηση της συμπεριφοράς του καταναλωτή μπορεί να οδηγήσει σε σημαντικές πληροφορίες για τους λιανέμπορους και να ενισχύσει την εμπειρία των καταναλωτών μέσα στο κατάστημα. Παρά τη δημοτικότητα της Χωρικής Αναλυτικής και των υπηρεσιών βάσει τοποθεσίας, και οι δύο αντιμετωπίζουν ένα κοινό πρόβλημα, την ακρίβεια του μηχανισμού εντοπισμού που χρησιμοποιούν. Για την αντιμετώπιση των ερευνητικών ερωτημάτων της παρούσας διατριβής, (α) υιοθετούμε ως μεθοδολογική προσέγγιση το μοντέλο Design Science, (β) αναπτύσσουμε ένα σύστημα εντοπισμού θέσης για εσωτερικά περιβάλλοντα και (γ) υιοθετούμε μια προσέγγιση μηχανικής μάθησης που εκτελεί εντοπισμό θέσης. Συνοπτικά, σχεδιάζουμε και αναπτύσσουμε ένα σύστημα που παράγει και επεξεργάζεται δεδομένα σήματος και αναπτύσσουμε μια προσέγγιση μηχανικής μάθησης για τον εντοπισμό θέσης που μπορεί να εφαρμοστεί σε χωροχρονικά δεδομένα από συσκευές Διαδικτύου των Πραγμάτων (IoT). Αξιολογούμε δύο διαφορετικές ασύρματες τεχνολογίες ((α) Wi-Fi και (β) Bluetooth Low Energy Beacons) και εφαρμόζουμε τεχνικές τεχνητής νοημοσύνης για την αντιμετώπιση των ερευνητικών ερωτημάτων. Επιπλέον, για την αντιμετώπιση των ερευνητικών ζητημάτων προτείνουμε ένα artifact συστήματος το οποίο είναι υπεύθυνο για τη δημιουργία, την καταγραφή και την επεξεργασία των δεδομένων για τον εντοπισμό θέσης σε εσωτερικά περιβάλλοντα. Το αποτέλεσμα της προτεινόμενης προσέγγισης είναι η θέση του χρήστη της ασύρματης υποδομής μέσα στο κατάστημα. Εφαρμόζουμε αυτήν την προσέγγιση σε δύο διαφορετικές περιπτώσεις. Η πρώτη περίπτωση αφορά την τεχνολογία BLE Beacons, ενώ η δεύτερη αφορά τεχνολογία Wi-Fi. Στη συνέχεια αξιολογούμε τα ευρήματα χρησιμοποιώντας τεχνική αποτίμηση της επίδοσης της και επίσης εξετάζουμε την επιχειρηματική ερμηνεία των αποτελεσμάτων. Για το σκοπό αυτό, αξιοποιούμε την αξιολόγηση βάσει δεδομένων και την αξιολόγηση βάσει χρηστών, προκειμένου να αξιολογήσουμε τα αποτελέσματα της προσέγγισης εντοπισμού θέσης. Τέλος, παραθέτουμε μια σειρά πρακτικές εφαρμογές με βάση τη διάσταση της θέσης.

show abstract

Unsupervised labelling of sequential data for location identification in indoor environments

Cited by 5 publications

References 42 publications

An Ensemble Filter for Indoor Positioning in a Retail Store Using Bluetooth Low Energy Beacons

An Ensemble Filter for Indoor Positioning in a Retail Store Using Bluetooth Low Energy Beacons

Out of hours workload management: Bayesian inference for decision support in secondary care

Indoor localization systems for retail stores

Contact Info

Product

Resources

About