K-MDTSC: K-Multi-Dimensional Time-Series Clustering Algorithm
The increasing capability to collect data gives us the possibility to collect a massive amount of heterogeneous data. Among the heterogeneous data available, time-series represents a mother lode of information yet to be fully explored. Current data mining techniques have several shortcomings while a...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2021-05-01
|
| Series: | Electronics |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2079-9292/10/10/1166 |
| _version_ | 1797534194572722176 |
|---|---|
| author | Danilo Giordano Marco Mellia Tania Cerquitelli |
| author_facet | Danilo Giordano Marco Mellia Tania Cerquitelli |
| author_sort | Danilo Giordano |
| collection | DOAJ |
| description | The increasing capability to collect data gives us the possibility to collect a massive amount of heterogeneous data. Among the heterogeneous data available, time-series represents a mother lode of information yet to be fully explored. Current data mining techniques have several shortcomings while analyzing time-series, especially when more than one time-series, i.e., multi-dimensional timeseries, should be analyzed together to extract knowledge from the data. In this context, we present <i>K-MDTSC</i> (K-Multi-Dimensional Time-Series Clustering), a novel clustering algorithm specifically designed to deal with multi-dimensional time-series. Firstly, we demonstrate <i>K-MDTSC</i> capability to group multi-dimensional time-series using synthetic datasets. We compare <i>K-MDTSC</i> results with <i>k-Shape</i>, a state-of-art time-series clustering algorithm based on K-means. Our results show both <i>K-MDTSC</i> and <i>k-Shape</i> create good clustering results. However, <i>K-MDTSC</i> outperforms <i>k-Shape</i> when complicating the synthetic dataset. Secondly, we apply <i>K-MDTSC</i> in a real case scenario where we are asked to replace a scheduled maintenance with a predictive approach. To this end, we create a generalized pipeline to process data from a real industrial plant welding process. We apply <i>K-MDTSC</i> to create clusters of weldings based on their welding shape. Our results show that <i>K-MDTSC</i> identifies different welding profiles, but that the aging of the electrode does not negatively impact the welding process. |
| first_indexed | 2024-03-10T11:27:08Z |
| format | Article |
| id | doaj.art-bccf494a0e8e45f4b7e059f0210fa5f3 |
| institution | Directory Open Access Journal |
| issn | 2079-9292 |
| language | English |
| last_indexed | 2024-03-10T11:27:08Z |
| publishDate | 2021-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Electronics |
| spelling | doaj.art-bccf494a0e8e45f4b7e059f0210fa5f32023-11-21T19:36:50ZengMDPI AGElectronics2079-92922021-05-011010116610.3390/electronics10101166K-MDTSC: K-Multi-Dimensional Time-Series Clustering AlgorithmDanilo Giordano0Marco Mellia1Tania Cerquitelli2Department of Control and Computer Engineering, Politecnico di Torino, Corso Duca degli Abruzzi, 24-10129 Turin, ItalyDepartment of Electronics and Telecommunications, Politecnico di Torino, Corso Duca degli Abruzzi, 24-10129 Turin, ItalyDepartment of Control and Computer Engineering, Politecnico di Torino, Corso Duca degli Abruzzi, 24-10129 Turin, ItalyThe increasing capability to collect data gives us the possibility to collect a massive amount of heterogeneous data. Among the heterogeneous data available, time-series represents a mother lode of information yet to be fully explored. Current data mining techniques have several shortcomings while analyzing time-series, especially when more than one time-series, i.e., multi-dimensional timeseries, should be analyzed together to extract knowledge from the data. In this context, we present <i>K-MDTSC</i> (K-Multi-Dimensional Time-Series Clustering), a novel clustering algorithm specifically designed to deal with multi-dimensional time-series. Firstly, we demonstrate <i>K-MDTSC</i> capability to group multi-dimensional time-series using synthetic datasets. We compare <i>K-MDTSC</i> results with <i>k-Shape</i>, a state-of-art time-series clustering algorithm based on K-means. Our results show both <i>K-MDTSC</i> and <i>k-Shape</i> create good clustering results. However, <i>K-MDTSC</i> outperforms <i>k-Shape</i> when complicating the synthetic dataset. Secondly, we apply <i>K-MDTSC</i> in a real case scenario where we are asked to replace a scheduled maintenance with a predictive approach. To this end, we create a generalized pipeline to process data from a real industrial plant welding process. We apply <i>K-MDTSC</i> to create clusters of weldings based on their welding shape. Our results show that <i>K-MDTSC</i> identifies different welding profiles, but that the aging of the electrode does not negatively impact the welding process.https://www.mdpi.com/2079-9292/10/10/1166machine learningclusteringpredictive maintenanceindustry 4.0 |
| spellingShingle | Danilo Giordano Marco Mellia Tania Cerquitelli K-MDTSC: K-Multi-Dimensional Time-Series Clustering Algorithm Electronics machine learning clustering predictive maintenance industry 4.0 |
| title | K-MDTSC: K-Multi-Dimensional Time-Series Clustering Algorithm |
| title_full | K-MDTSC: K-Multi-Dimensional Time-Series Clustering Algorithm |
| title_fullStr | K-MDTSC: K-Multi-Dimensional Time-Series Clustering Algorithm |
| title_full_unstemmed | K-MDTSC: K-Multi-Dimensional Time-Series Clustering Algorithm |
| title_short | K-MDTSC: K-Multi-Dimensional Time-Series Clustering Algorithm |
| title_sort | k mdtsc k multi dimensional time series clustering algorithm |
| topic | machine learning clustering predictive maintenance industry 4.0 |
| url | https://www.mdpi.com/2079-9292/10/10/1166 |
| work_keys_str_mv | AT danilogiordano kmdtsckmultidimensionaltimeseriesclusteringalgorithm AT marcomellia kmdtsckmultidimensionaltimeseriesclusteringalgorithm AT taniacerquitelli kmdtsckmultidimensionaltimeseriesclusteringalgorithm |