hh.sePublications
Change search
Link to record
Permanent link

Direct link
BETA
Alternative names
Publications (10 of 56) Show all publications
Sheikholharam Mashhadi, P., Nowaczyk, S. & Pashami, S. (2020). Stacked Ensemble of Recurrent Neural Networks for Predicting Turbocharger Remaining Useful Life. Applied Sciences, 10(1)
Open this publication in new window or tab >>Stacked Ensemble of Recurrent Neural Networks for Predicting Turbocharger Remaining Useful Life
2020 (English)In: Applied Sciences, E-ISSN 2076-3417, Vol. 10, no 1Article in journal (Refereed) Epub ahead of print
Abstract [en]

Predictive Maintenance (PM) is a proactive maintenance strategy that tries to minimize a system’s downtime by predicting failures before they happen. It uses data from sensors to measure the component’s state of health and make forecasts about its future degradation. However, existing PM methods typically focus on individual measurements. While it is natural to assume that a history of measurements carries more information than a single one. This paper aims at incorporating such information into PM models. In practice, especially in the automotive domain, diagnostic models have low performance, due to a large amount of noise in the data and limited sensing capability. To address this issue, this paper proposes to use a specific type of ensemble learning known as Stacked Ensemble. The idea is to aggregate predictions of multiple models—consisting of Long Short-Term Memory (LSTM) and Convolutional-LSTM—via a meta model, in order to boost performance. Stacked Ensemble model performs well when its base models are as diverse as possible. To this end, each such model is trained using a specific combination of the following three aspects: feature subsets, past dependency horizon, and model architectures. Experimental results demonstrate benefits of the proposed approach on a case study of heavy-duty truck turbochargers. © 2019 MDPI

Place, publisher, year, edition, pages
Basel: MDPI, 2020
Keywords
predictive maintenance, remaining useful life, recurrent neural networks, LSTM, Stacked Ensemble
National Category
Other Computer and Information Science
Identifiers
urn:nbn:se:hh:diva-41314 (URN)10.3390/app10010069 (DOI)
Projects
HEALTH-VINNOVA
Funder
Vinnova
Available from: 2019-12-29 Created: 2019-12-29 Last updated: 2019-12-30
Dahl, O., Johansson, F., Khoshkangini, R., Pashami, S., Nowaczyk, S. & Pihl, C. (2020). Understanding Association Between Logged Vehicle Data and Vehicle Marketing Parameters -- Using Clustering and Rule-Based Machine Learning. In: : . Paper presented at The 3rd International Conference on Information Management and Processing (ICIMP 2020), Portsmouth, UK, January 9-11, 2020.
Open this publication in new window or tab >>Understanding Association Between Logged Vehicle Data and Vehicle Marketing Parameters -- Using Clustering and Rule-Based Machine Learning
Show others...
2020 (English)Conference paper, Published paper (Refereed)
Abstract [en]

Trucks are designed, configured and marketed for various working environments. There lies a concern whether trucks are used as intended by the manufacturer, as usage may impact the longevity, efficiency and productivity of the trucks.

In this paper we propose a framework that aims to extract costumers' vehicle behaviours from LVD in order to evaluate whether they align with vehicle configurations, so-called GTA parameters. GMMs are employed to cluster and classify various vehicle behaviors from the LVD. RBML was applied on the clusters to examine whether vehicle behaviors follow the GTA configuration. Particularly, we propose an approach based on studying associations that is able to extract insights on whether the trucks are used as intended. Experimental results shown that while for the vast majority of the trucks' behaviors seemingly follows their GTA configuration, there are also interesting outliers that warrant further analysis.

Keywords
Machine Learning, Clustering, Usage Behaviors, Association Rule Mining, Gaussian Mixture Models.
National Category
Computer Sciences
Identifiers
urn:nbn:se:hh:diva-41214 (URN)
Conference
The 3rd International Conference on Information Management and Processing (ICIMP 2020), Portsmouth, UK, January 9-11, 2020
Available from: 2019-12-10 Created: 2019-12-10 Last updated: 2019-12-11
Calikus, E., Nowaczyk, S., Pinheiro Sant'Anna, A., Gadd, H. & Werner, S. (2019). A data-driven approach for discovering heat load patterns in district heating. Applied Energy, 252, Article ID 113409.
Open this publication in new window or tab >>A data-driven approach for discovering heat load patterns in district heating
Show others...
2019 (English)In: Applied Energy, ISSN 0306-2619, E-ISSN 1872-9118, Vol. 252, article id 113409Article in journal (Refereed) Published
Abstract [en]

Understanding the heat usage of customers is crucial for effective district heating operations and management. Unfortunately, existing knowledge about customers and their heat load behaviors is quite scarce. Most previous studies are limited to small-scale analyses that are not representative enough to understand the behavior of the overall network. In this work, we propose a data-driven approach that enables large-scale automatic analysis of heat load patterns in district heating networks without requiring prior knowledge. Our method clusters the customer profiles into different groups, extracts their representative patterns, and detects unusual customers whose profiles deviate significantly from the rest of their group. Using our approach, we present the first large-scale, comprehensive analysis of the heat load patterns by conducting a case study on many buildings in six different customer categories connected to two district heating networks in the south of Sweden. The 1222 buildings had a total floor space of 3.4 million square meters and used 1540 TJ heat during 2016. The results show that the proposed method has a high potential to be deployed and used in practice to analyze and understand customers’ heat-use habits. © 2019 Calikus et al. Published by Elsevier Ltd.

Place, publisher, year, edition, pages
Oxford: Elsevier, 2019
Keywords
District heating, Energy efficiency, Heat load patterns, Clustering, Abnormal heat use
National Category
Other Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
urn:nbn:se:hh:diva-40907 (URN)10.1016/j.apenergy.2019.113409 (DOI)000497968000013 ()2-s2.0-85066961984 (Scopus ID)
Funder
Knowledge Foundation, 20160103
Available from: 2019-11-12 Created: 2019-11-12 Last updated: 2019-12-11Bibliographically approved
Khan, T., Lundgren, L., Anderson, D. G., Novak, I., Dougherty, M., Verikas, A., . . . Aharonson, V. (2019). Assessing Parkinson's disease severity using speech analysis in non-native speakers. Computer speech & language (Print), Article ID 101047.
Open this publication in new window or tab >>Assessing Parkinson's disease severity using speech analysis in non-native speakers
Show others...
2019 (English)In: Computer speech & language (Print), ISSN 0885-2308, E-ISSN 1095-8363, article id 101047Article in journal (Refereed) Published
Abstract [en]

Background: Speech disorder is a common manifestation of Parkinson's disease with two main symptoms, dysprosody and dysphonia. Previous research studying objective measures of speech symptoms involved patients and examiners who were native language speakers. Measures such as cepstral separation difference (CSD) features to quantify dysphonia and dysprosody accurately distinguish the severity of speech impairment. Importantly CSD, together with other speech features, including Mel-frequency coefficients, fundamental-frequency variation, and spectral dynamics, characterize speech intelligibility in PD. However, non-native language speakers transfer phonological rules of their mother language that tamper speech assessment.

Objectives: This paper explores CSD's capability: first, to quantify dysprosody and dysphonia of non-native language speakers, Parkinson patients and controls, and secondly, to characterize the severity of speech impairment when Parkinson's dysprosody accompanies non-native linguistic dysprosody.

Methods: CSD features were extracted from 168 speech samples recorded from 19 healthy controls, 15 rehabilitated and 23 not-rehabilitated Parkinson patients in three different clinical speech tests based on Unified Parkinson's disease rating scale motor-speech examination. Statistical analyses were performed to compare groups using analysis of variance, intraclass correlation, and Guttman correlation coefficient µ2. Random forests were trained to classify the severity of speech impairment using CSD and the other speech features. Feature importance in classification was determined using permutation importance score.

Results: Results showed that the CSD feature describing dysphonia was uninfluenced by non-native accents, strongly correlated with the clinical examination (µ2>0.5), and significantly discriminated between the healthy, rehabilitated, and not-rehabilitated patient groups based on the severity of speech symptoms. However, the feature describing dysprosody did not correlate with the clinical examination but significantly distinguished the groups. The classification model based on random forests and selected features characterized the severity of speech impairment of non-native language speakers with high accuracy. Importantly, the permutation importance score of the CSD feature representing dysphonia was the highest compared to other features. Results showed a strong negative correlation (µ2<-0.5) between L-dopa administration and the CSD features.

Conclusions: Although non-native accents reduce speech intelligibility, the CSD features can accurately characterize speech impairment, which is not always possible in the clinical examination. Findings support using CSD for monitoring Parkinson's disease.

© 2019 Elsevier Ltd. All rights reserved.

Place, publisher, year, edition, pages
London, UK: Academic Press, 2019
Keywords
Dysphonia, Dysprosody, Parkinson's disease, Speech processing, Tele-monitoring
National Category
Language Technology (Computational Linguistics)
Identifiers
urn:nbn:se:hh:diva-41003 (URN)10.1016/j.csl.2019.101047 (DOI)
Available from: 2019-11-21 Created: 2019-11-21 Last updated: 2019-11-27Bibliographically approved
Pirasteh, P., Nowaczyk, S., Pashami, S., Löwenadler, M., Thunberg, K., Ydreskog, H. & Berck, P. (2019). Interactive feature extraction for diagnostic trouble codes in predictive maintenance: A case study from automotive domain. In: Proceedings of the Workshop on Interactive Data Mining: . Paper presented at WSDM 2019: The 12th ACM International Conference on Web Search and Data Mining, Melbourne, VIC, Australia, 11-15 February, 2019. New York, NY: Association for Computing Machinery (ACM), Article ID 4.
Open this publication in new window or tab >>Interactive feature extraction for diagnostic trouble codes in predictive maintenance: A case study from automotive domain
Show others...
2019 (English)In: Proceedings of the Workshop on Interactive Data Mining, New York, NY: Association for Computing Machinery (ACM), 2019, article id 4Conference paper, Published paper (Refereed)
Abstract [en]

Predicting future maintenance needs of equipment can be addressed in a variety of ways. Methods based on machine learning approaches provide an interesting platform for mining large data sets to find patterns that might correlate with a given fault. In this paper, we approach predictive maintenance as a classification problem and use Random Forest to separate data readouts within a particular time window into those corresponding to faulty and non-faulty component categories. We utilize diagnostic trouble codes (DTCs) as an example of event-based data, and propose four categories of features that can be derived from DTCs as a predictive maintenance framework. We test the approach using large-scale data from a fleet of heavy duty trucks, and show that DTCs can be used within our framework as indicators of imminent failures in different components.

Place, publisher, year, edition, pages
New York, NY: Association for Computing Machinery (ACM), 2019
Keywords
Predictive maintenance, failure detection, diagnostic trouble codes, feature extraction
National Category
Signal Processing
Identifiers
urn:nbn:se:hh:diva-40184 (URN)10.1145/3304079.3310288 (DOI)978-1-4503-6296-2 (ISBN)
Conference
WSDM 2019: The 12th ACM International Conference on Web Search and Data Mining, Melbourne, VIC, Australia, 11-15 February, 2019
Available from: 2019-07-07 Created: 2019-07-07 Last updated: 2019-08-02Bibliographically approved
Calikus, E., Fan, Y., Nowaczyk, S. & Pinheiro Sant'Anna, A. (2019). Interactive-cosmo: Consensus self-organized models for fault detection with expert feedback. In: Proceedings of the Workshop on Interactive Data Mining, WIDM 2019: . Paper presented at 1st Workshop on Interactive Data Mining, WIDM 2019, co-located with 12th ACM International Conference on Web Search and Data Mining, WSDM 2019, Melbourne, Australia; 15 February, 2019 (pp. 1-9). New York: Association for Computing Machinery (ACM)
Open this publication in new window or tab >>Interactive-cosmo: Consensus self-organized models for fault detection with expert feedback
2019 (English)In: Proceedings of the Workshop on Interactive Data Mining, WIDM 2019, New York: Association for Computing Machinery (ACM), 2019, p. 1-9Conference paper, Published paper (Refereed)
Abstract [en]

Diagnosing deviations and predicting faults is an important task, especially given recent advances related to Internet of Things. However, the majority of the efforts for diagnostics are still carried out by human experts in a time-consuming and expensive manner. One promising approach towards self-monitoring systems is based on the "wisdom of the crowd" idea, where malfunctioning equipments are detected by understanding the similarities and differences in the operation of several alike systems.

A fully autonomous fault detection, however, is not possible, since not all deviations or anomalies correspond to faulty behaviors; many can be explained by atypical usage or varying external conditions. In this work, we propose a method which gradually incorporates expert-provided feedback for more accurate self-monitoring. Our idea is to support model adaptation while allowing human feedback to persist over changes in data distribution, such as concept drift. © 2019 Association for Computing Machinery.

Place, publisher, year, edition, pages
New York: Association for Computing Machinery (ACM), 2019
Keywords
Anomaly Detection, Self-Monitoring, Active Learning, Human-in- the-loop
National Category
Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
urn:nbn:se:hh:diva-41365 (URN)10.1145/3304079.3310289 (DOI)2-s2.0-85069779014 (Scopus ID)978-1-4503-6296-2 (ISBN)
Conference
1st Workshop on Interactive Data Mining, WIDM 2019, co-located with 12th ACM International Conference on Web Search and Data Mining, WSDM 2019, Melbourne, Australia; 15 February, 2019
Available from: 2020-01-10 Created: 2020-01-10 Last updated: 2020-01-14Bibliographically approved
Ashfaq, A. & Nowaczyk, S. (2019). Machine learning in healthcare - a system’s perspective. In: B. Aditya Prakash, Anil Vullikanti, Shweta Bansal, Adam Sadelik (Ed.), Proceedings of the ACM SIGKDD Workshop on Epidemiology meets Data Mining and Knowledge Discovery (epiDAMIK): . Paper presented at 25th ACM SIGKDD Workshop on Epidemiology meets Data Mining and Knowledge Discovery (epiDAMIK '19), Anchorage, Alaska, United States, August 5, 2019 (pp. 14-17). Arlington
Open this publication in new window or tab >>Machine learning in healthcare - a system’s perspective
2019 (English)In: Proceedings of the ACM SIGKDD Workshop on Epidemiology meets Data Mining and Knowledge Discovery (epiDAMIK) / [ed] B. Aditya Prakash, Anil Vullikanti, Shweta Bansal, Adam Sadelik, Arlington, 2019, p. 14-17Conference paper, Published paper (Refereed)
Abstract [en]

A consequence of the fragmented and siloed healthcare landscape is that patient care (and data) is split along multitude of different facilities and computer systems and enabling interoperability between these systems is hard. The lack interoperability not only hinders continuity of care and burdens providers, but also hinders effective application of Machine Learning (ML) algorithms. Thus, most current ML algorithms, designed to understand patient care and facilitate clinical decision-support, are trained on limited datasets. This approach is analogous to the Newtonian paradigm of Reductionism in which a system is broken down into elementary components and a description of the whole is formed by understanding those components individually. A key limitation of the reductionist approach is that it ignores the component-component interactions and dynamics within the system which are often of prime significance in understanding the overall behaviour of complex adaptive systems (CAS). Healthcare is a CAS.

Though the application of ML on health data have shown incremental improvements for clinical decision support, ML has a much a broader potential to restructure care delivery as a whole and maximize care value. However, this ML potential remains largely untapped: primarily due to functional limitations of Electronic Health Records (EHR) and the inability to see the healthcare system as a whole. This viewpoint (i) articulates the healthcare as a complex system which has a biological and an organizational perspective, (ii) motivates with examples, the need of a system's approach when addressing healthcare challenges via ML and, (iii) emphasizes to unleash EHR functionality - while duly respecting all ethical and legal concerns - to reap full benefits of ML.

Place, publisher, year, edition, pages
Arlington: , 2019
Keywords
Machine learning, Healthcare complexity, System's thinking, Electronic health records
National Category
Other Medical Engineering
Identifiers
urn:nbn:se:hh:diva-40395 (URN)
Conference
25th ACM SIGKDD Workshop on Epidemiology meets Data Mining and Knowledge Discovery (epiDAMIK '19), Anchorage, Alaska, United States, August 5, 2019
Available from: 2019-08-14 Created: 2019-08-14 Last updated: 2019-08-14Bibliographically approved
Chen, K., Pashami, S., Fan, Y. & Nowaczyk, S. (2019). Predicting Air Compressor Failures Using Long Short Term Memory Networks. In: Paulo Moura Oliveira, Paulo Novais, Luís Paulo Reis (Ed.), Progress in Artificial Intelligence: 19th EPIA Conference on Artificial Intelligence, EPIA 2019, Vila Real, Portugal, September 3–6, 2019, Proceedings, Part I. Paper presented at 19th EPIA Conference on Artificial Intelligence (EPIA 2019), Vila Real, Portugal, September 3–6, 2019 (pp. 596-609). Cham: Springer
Open this publication in new window or tab >>Predicting Air Compressor Failures Using Long Short Term Memory Networks
2019 (English)In: Progress in Artificial Intelligence: 19th EPIA Conference on Artificial Intelligence, EPIA 2019, Vila Real, Portugal, September 3–6, 2019, Proceedings, Part I / [ed] Paulo Moura Oliveira, Paulo Novais, Luís Paulo Reis, Cham: Springer, 2019, p. 596-609Conference paper, Published paper (Refereed)
Abstract [en]

We introduce an LSTM-based method for predicting compressor failures using aggregated sensory data, and evaluate it using historical information from over 1000 heavy duty vehicles during 2015 and 2016. The goal is to proactively identify trucks that will require maintenance in the near future, so that component replacement can be scheduled before the failure happens, translating into improved uptime. The problem is formulated as a classification task of whether a compressor failure will happen within the specified prediction horizon. A recurrent neural network using Long Short-Term Memory (LSTM) architecture is employed as the prediction model, and compared against Random Forest (RF), the solution used in industrial deployment at the moment. Experimental results show that while Random Forest slightly outperforms LSTM in terms of AUC score, the predictions of LSTM stay significantly more stable over time, showing a consistent trend from healthy to faulty class. Additionally, LSTM is also better at detecting the switch from faulty class to the healthy one after a repair. We demonstrate that this stability is important for making repair decisions, especially in questionable cases, and therefore LSTM model is likely to lead to better results in practice. © Springer Nature Switzerland AG 2019

Place, publisher, year, edition, pages
Cham: Springer, 2019
Series
Lecture Notes in Computer Science, ISSN 0302-9743 ; 11804
Keywords
Fault detection, Predictive maintenance, Recurrent neural networks, Long-short term memory
National Category
Computer Systems
Identifiers
urn:nbn:se:hh:diva-41366 (URN)10.1007/978-3-030-30241-2_50 (DOI)2-s2.0-85072895300 (Scopus ID)978-3-030-30240-5 (ISBN)978-3-030-30241-2 (ISBN)
Conference
19th EPIA Conference on Artificial Intelligence (EPIA 2019), Vila Real, Portugal, September 3–6, 2019
Available from: 2020-01-10 Created: 2020-01-10 Last updated: 2020-01-14Bibliographically approved
Ashfaq, A., Pinheiro Sant'Anna, A., Lingman, M. & Nowaczyk, S. (2019). Readmission prediction using deep learning on electronic health records. Journal of Biomedical Informatics, 97, Article ID 103256.
Open this publication in new window or tab >>Readmission prediction using deep learning on electronic health records
2019 (English)In: Journal of Biomedical Informatics, ISSN 1532-0464, E-ISSN 1532-0480, Vol. 97, article id 103256Article in journal (Refereed) Published
Abstract [en]

Unscheduled 30-day readmissions are a hallmark of Congestive Heart Failure (CHF) patients that pose significant health risks and escalate care cost. In order to reduce readmissions and curb the cost of care, it is important to initiate targeted intervention programs for patients at risk of readmission. This requires identifying high-risk patients at the time of discharge from hospital. Here, using real data from over 7,500 CHF patients hospitalized between 2012 and 2016 in Sweden, we built and tested a deep learning framework to predict 30-day unscheduled readmission. We present a cost-sensitive formulation of Long Short-Term Memory (LSTM) neural network using expert features and contextual embedding of clinical concepts. This study targets key elements of an Electronic Health Record (EHR) driven prediction model in a single framework: using both expert and machine derived features, incorporating sequential patterns and addressing the class imbalance problem. We show that the model with all key elements achieves a higher discrimination ability (AUC 0.77) compared to the rest. Additionally, we present a simple financial analysis to estimate annual savings if targeted interventions are offered to high risk patients. © 2019 The Authors

Place, publisher, year, edition, pages
Maryland Heights, MO: Academic Press, 2019
Keywords
Electronic health records, Readmission prediction, Long short-term memory networks, Contextual embeddings
National Category
Health Care Service and Management, Health Policy and Services and Health Economy
Identifiers
urn:nbn:se:hh:diva-39297 (URN)10.1016/j.jbi.2019.103256 (DOI)31351136 (PubMedID)2-s2.0-85069858722 (Scopus ID)
Projects
HiCube - behovsmotiverad hälsoinnovation
Funder
European Regional Development Fund (ERDF)
Note

Funding: The authors thank the European Regional Development Fund (ERDF), Health Technology Center and CAISR at Halmstad University and Hallands Hospital for financing the research work under the project HiCube - behovsmotiverad hälsoinnovation.

Available from: 2019-04-30 Created: 2019-04-30 Last updated: 2019-09-10Bibliographically approved
Mashad Nemati, H., Pinheiro Sant'Anna, A., Nowaczyk, S., Jürgensen, J. H. & Hilber, P. (2019). Reliability Evaluation of Power Cables Considering the Restoration Characteristic. International Journal of Electrical Power & Energy Systems, 105, 622-631
Open this publication in new window or tab >>Reliability Evaluation of Power Cables Considering the Restoration Characteristic
Show others...
2019 (English)In: International Journal of Electrical Power & Energy Systems, ISSN 0142-0615, E-ISSN 1879-3517, Vol. 105, p. 622-631Article in journal (Refereed) Published
Abstract [en]

In this paper Weibull parametric proportional hazard model (PHM) is used to estimate the failure rate of every individual cable based on its age and a set of explanatory factors. The required information for the proposed method is obtained by exploiting available historical cable inventory and failure data. This data-driven method does not require any additional measurements on the cables, and allows the cables to be ranked for maintenance prioritization and repair actions.

Furthermore, the results of reliability analysis of power cables are compared when the cables are considered as repairable or non-repairable components. The paper demonstrates that the methods which estimate the time-to-the-first failure (for non-repairable components) lead to incorrect conclusions about reliability of repairable power cables.

The proposed method is used to evaluate the failure rate of each individual Paper Insulated Lead Cover (PILC) underground cables in a distribution grid in the south of Sweden. © 2018 Elsevier Ltd

Place, publisher, year, edition, pages
London: Elsevier, 2019
Keywords
Power cable, historical data, reliability, proportional hazard model, preventive maintenance.
National Category
Electrical Engineering, Electronic Engineering, Information Engineering Other Electrical Engineering, Electronic Engineering, Information Engineering
Identifiers
urn:nbn:se:hh:diva-35470 (URN)10.1016/j.ijepes.2018.08.047 (DOI)000449447200055 ()2-s2.0-85053080255 (Scopus ID)
Available from: 2017-11-24 Created: 2017-11-24 Last updated: 2019-03-19Bibliographically approved
Projects
iMedA: Improving MEDication Adherence through Person Centered Care and Adaptive Interventions [2017-04617_Vinnova]; Halmstad University
Organisations
Identifiers
ORCID iD: ORCID iD iconorcid.org/0000-0002-7796-5201

Search in DiVA

Show all publications