803 Hits in 5.6 sec

Rethinking 1D-CNN for Time Series Classification: A Stronger Baseline [article]

Wensi Tang, Guodong Long, Lu Liu, Tianyi Zhou, Jing Jiang, Michael Blumenstein
2021 arXiv   pre-print
For time series classification task using 1D-CNN, the selection of kernel size is critically important to ensure the model can capture the right scale salient signal from a long time-series.  ...  We also published the experimental source codes at GitHub (  ...  Introduction For TSC tasks, the optimal feature extraction scales usually vary across datasets [Dau et al., 2018] .  ... 
arXiv:2002.10061v2 fatcat:4n77dgwuu5gdhmrrqaxvaq57nm

The Potential of Solar Air Heating in the Turkish Industrial Sector

Balázs Bokor, Hacer Akhan, Dogan Eryener, László Kajtár
2018 Periodica Polytechnica: Mechanical Engineering  
Generally, it can be stated that a location with cold climate and high solar radiation at the same time benefits most from the use of a TSC system.  ...  In this paper, the effect of different climatic zones on the thermal performance of the TSC is investigated.  ...  Acknowledgement The international cooperation between Trakya University and Budapest University of Technology and Economics has been funded by the "National Talent Program" of the Hungarian Ministry of  ... 
doi:10.3311/ppme.13028 fatcat:wd2ec5qslvbanaxxqud4lppmma

Interpretable Time Series Classification using Linear Models and Multi-resolution Multi-domain Symbolic Representations [article]

Thach Le Nguyen and Severin Gsponer and Iulia Ilie and Martin O'Reilly and Georgiana Ifrim
2020 arXiv   pre-print
accurate as deep learning models but are more efficient regarding running time and memory, can work with variable-length time series and can be interpreted by highlighting the discriminative symbolic features  ...  This aspect of classifiers has become critical for many application domains and the introduction of the EU GDPR legislation in 2018 is likely to further emphasize the importance of interpretable learning  ...  We would also like to gratefully acknowledge the work by researchers at University of California Riverside, USA (especially Eamonn Keogh and his team) and researchers at University of East Anglia, UK (  ... 
arXiv:2006.01667v1 fatcat:wrgj77kmzbdlvhenrfqz47jgjm

DSCo: A Language Modeling Approach for Time Series Classification [chapter]

Daoyuan Li, Li Li, Tegawendé F. Bissyandé, Jacques Klein, Yves Le Traon
2016 Lecture Notes in Computer Science  
Our work innovatively takes advantage of mature techniques from both time series mining and NLP communities.  ...  Symbolic representation of time series -which transforms numeric time series data into texts -is a promising technique to address these challenges.  ...  Fig. 3 : 3 Illustration of DSCo's classification process. Fig. 4 : 4 1NN classification accuracy comparison between DTW (dashed) and SAX (solid) distance.  ... 
doi:10.1007/978-3-319-41920-6_22 fatcat:jtherbmd35czxnxezlyllbhavy

Deep learning for time series classification: a review

Hassan Ismail Fawaz, Germain Forestier, Jonathan Weber, Lhassane Idoumghar, Pierre-Alain Muller
2019 Data mining and knowledge discovery  
In this article, we study the current state-of-the-art performance of deep learning algorithms for TSC by presenting an empirical study of the most recent DNN architectures for TSC.  ...  Time Series Classification (TSC) is an important and challenging problem in data mining. With the increase of time series data availability, hundreds of TSC algorithms have been proposed.  ...  This work was supported by the ANR TIMES project (grant ANR-17-CE23-0015) of the French Agence Nationale de la Recherche.  ... 
doi:10.1007/s10618-019-00619-1 fatcat:n4mw55rqxvf6pobw4okh6t7744

Squeezed Very Deep Convolutional Neural Networks for Text Classification [article]

Andréa B. Duque, Luã Lázaro J. Santos, David Macêdo, Cleber Zanchettin
2019 arXiv   pre-print
Most of the research in convolutional neural networks has focused on increasing network depth to improve accuracy, resulting in a massive number of parameters which restricts the trained network to platforms  ...  The squeezed model (SVDCNN) is between 10x and 20x smaller, depending on the network depth, maintaining a maximum size of 6MB.  ...  Convolution with kernel size 3 and 256 feature maps as output.  ... 
arXiv:1901.09821v1 fatcat:lxjx73xcfrfrhddesybv7i4dgm

DSCo-NG: A Practical Language Modeling Approach for Time Series Classification [chapter]

Daoyuan Li, Tegawendé F. Bissyandé, Jacques Klein, Yves Le Traon
2016 Lecture Notes in Computer Science  
However, this approach was flawed in practice due to its excessive memory usage and the need for a priori knowledge about the dataset.  ...  The abundance of time series data in various domains and their high dimensionality characteristic are challenging for harvesting useful information from them.  ...  Acknowledgment The authors would like to thank Paul Wurth S.A. and Luxembourg Ministry of Economy for sponsoring this research work.  ... 
doi:10.1007/978-3-319-46349-0_1 fatcat:r74kcumiifhrlph3rrge4qqt7y

Deep Semi-Supervised Learning for Time Series Classification [article]

Jann Goschenhofer, Rasmus Hvingelby, David Rügamer, Janek Thomas, Moritz Wagner, Bernd Bischl
2021 arXiv   pre-print
We perform extensive comparisons under a decidedly realistic and appropriate evaluation scheme with a unified reimplementation of all algorithms considered, which is yet lacking in the field.  ...  with varying amounts of labelled samples.  ...  of BAYERN DIGITAL II (20-3410-2-9-8).  ... 
arXiv:2102.03622v1 fatcat:t7arhdtrw5bz3lum7r7wgufygu

Outlier Detection as Instance Selection Method for Feature Selection in Time Series Classification [article]

David Cemernek
2021 arXiv   pre-print
An important step in the preprocessing phase is feature selection, which aims at better performance of prediction models by reducing the amount of features of a data set.  ...  The aim of this work was to filter instances provided to feature selection methods for these rare instances, and thus positively influence the feature selection process.  ...  A possibly more intuitive comparison of traditional classification to TSC is based on the assumption that the former only uses static features, whereas TSC uses dynamic features, for which the change in  ... 
arXiv:2111.09127v1 fatcat:qxptgur6gjhdtfsgjagqk3oe7u

Transpired solar collectors in building service engineering: Combined system operation and special applications

B. Bokor, L. Kajtár
2018 International Review of Applied Sciences and Engineering  
Acknowledgements Supported by the ÚNKP-17-3-III New National Excellence Program of the Ministry of Human Capacities. Author B. Bokor hereby expresses thanks for the support.  ...  and about TSC and heat recovery comparison.  ...  During the comparison of TSC and heat recovery use in an air conditioning system all benefi t sources of the TSC have to be taken into consideration.  ... 
doi:10.1556/1848.2018.9.1.9 fatcat:eoksrgcg7ncxngeoiqmw4ry3wa

Traffic Signal Control via Reinforcement Learning for Reducing Global Vehicle Emission

Bálint Kővári, Lászlo Szőke, Tamás Bécsi, Szilárd Aradi, Péter Gáspár
2021 Sustainability  
The core of our solution is a novel rewarding concept for deep reinforcement learning (DRL) which does not utilize any reward shaping, hence exposes new insights into the traffic signal control (TSC) problem  ...  Moreover, the sustainability of the realized controls is also placed under investigation to evaluate their environmental impacts.  ...  A DQN and a PG agent are trained for solving the control task at hand.  ... 
doi:10.3390/su132011254 fatcat:mlg2pn55uncgxnum6bfbruiulm

High-fidelity promoter profiling reveals widespread alternative promoter usage and transposon-driven developmental gene expression

P. Batut, A. Dobin, C. Plessy, P. Carninci, T. R. Gingeras
2012 Genome Research  
developmental time-course of promoter usage.  ...  RAMPAGE features a streamlined protocol for fast and easy generation of highly multiplexed sequencing libraries, offers very high transcription start site specificity, generates accurate and reproducible  ...  Financial support was provided by the Watson School of Biological Sciences and the Florence Gould Foundation (P.B.).  ... 
doi:10.1101/gr.139618.112 pmid:22936248 pmcid:PMC3530677 fatcat:wkfn3v3tpvh4ljxipqlzcdkwfu

CLIR System Evaluation at the Second NTCIR Workshop [chapter]

Noriko Kando
2002 Lecture Notes in Computer Science  
A brief history, tasks, participants, test collections, CLIR evaluation at the workshops, and plan for the next workshop are described in this paper.  ...  This paper introduces , a series of evaluation workshops, which is designed to enhance research in information retrieval and related text processing techniques, such as summarization, extraction, by providing  ...  The CLIR task at the NTCIR Workshop 3 will be organized by the organizers of CHTR and JEIR, and HANTEC group.  ... 
doi:10.1007/3-540-45691-0_35 fatcat:e7hs7aldgncv3mqr74j3zolbqy

Computational results for an automatically tuned CMA-ES with increasing population size on the CEC'05 benchmark set

Tianjun Liao, Marco A. Montes de Oca, Thomas Stützle
2012 Soft Computing - A Fusion of Foundations, Methodologies and Applications  
Furthermore, we provide some further analyses on the impact of the modified parameter settings on iCMA-ES performance and a comparison with recent results of algorithms that use CMA-ES as a subordinate  ...  In particular, we consider a separation between tuning and test sets and, thus, tune iCMA-ES on a different set of functions than the ones of the CEC'05 benchmark set.  ...  We summarize the comparison with iCMA-ES-tsc in Table 7 . iCMA-ES-tsc performs statistically significantly better than iCMA-ES-05 on dimension 30 and reaches on more functions statistically significantly  ... 
doi:10.1007/s00500-012-0946-x fatcat:o6efxyujsjhujjhsvwdtzxxcly

InceptionTime: Finding AlexNet for time series classification

Hassan Ismail Fawaz, Benjamin Lucas, Germain Forestier, Charlotte Pelletier, Daniel F. Schmidt, Jonathan Weber, Geoffrey I. Webb, Lhassane Idoumghar, Pierre-Alain Muller, François Petitjean
2020 Data mining and knowledge discovery  
This paper brings deep learning at the forefront of research into time series classification (TSC). TSC is the area of machine learning tasked with the categorization (or labelling) of time series.  ...  Meanwhile deep learning has received enormous attention because of its high accuracy and scalability. Recent approaches to deep learning for TSC have been scalable, but less accurate than HIVE-COTE.  ...  This material is based upon work supported by the Air Force Office of Scientific Research, Asian Office of Aerospace Research and Development (AOARD) under award Number FA2386-18-1-4030.  ... 
doi:10.1007/s10618-020-00710-y fatcat:jlnn6zibyjh5xmwqzg7s2eniuq
« Previous Showing results 1 — 15 out of 803 results