dtcr
Learning Representations for Time Series Clustering
Qianli Ma, Jiawei Zheng, Sen Li, Gary W. Cottrell
Neural Information Processing Systems http://nips.cc/
Learning Representations for Time Series Clustering
Time series clustering is an essential unsupervised technique in cases when category information is not available. It has been widely applied to genome data, anomaly detection, and in general, in any domain where pattern detection is important. Although feature-based time series clustering methods are robust to noise and outliers, and can reduce the dimensionality of the data, they typically rely on domain knowledge to manually construct high-quality features. Sequence to sequence (seq2seq) models can learn representations from sequence data in an unsupervised manner by designing appropriate learning objectives, such as reconstruction and context prediction. When applying seq2seq to time series clustering, obtaining a representation that effectively represents the temporal dynamics of the sequence, multi-scale features, and good clustering properties remains a challenge.
Learning Representations for Time Series Clustering
Qianli Ma, Jiawei Zheng, Sen Li, Gary W. Cottrell
Neural Information Processing Systems http://nips.cc/
Reply to Reviewer # 1
Q1: What other ways to generate fake sequences may be suitable for this problem? A1: That is a good question. GAN to generate some more difficult fake sequences to further improve the ability of the encoder. Q1: Comparison with other state-of-the-art deep clustering methods which are not designed for time-series. A1: Following your suggestion, we compare our method with two state-of-the-art deep clustering methods (i.e., DEC (Xie et al., Table 1: Comparisons on 36 time series datasets (The No. of datasets is consistent with the one in Table 2 in main text)Dataset DEC(RI) IDEC(RI) DTCR(RI) DTCR(NMI) DTCR(ACC) Dataset DEC(RI) IDEC(RI) DTCR(RI) DTCR(NMI) DTCR(ACC)1 0.5817 0.6210 0.6868(0.0026)
Learning Representations for Time Series Clustering
Time series clustering is an essential unsupervised technique in cases when category information is not available. It has been widely applied to genome data, anomaly detection, and in general, in any domain where pattern detection is important. Although feature-based time series clustering methods are robust to noise and outliers, and can reduce the dimensionality of the data, they typically rely on domain knowledge to manually construct high-quality features. Sequence to sequence (seq2seq) models can learn representations from sequence data in an unsupervised manner by designing appropriate learning objectives, such as reconstruction and context prediction. When applying seq2seq to time series clustering, obtaining a representation that effectively represents the temporal dynamics of the sequence, multi-scale features, and good clustering properties remains a challenge.
Learning Representations for Time Series Clustering
Time series clustering is an essential unsupervised technique in cases when category information is not available. It has been widely applied to genome data, anomaly detection, and in general, in any domain where pattern detection is important. Although feature-based time series clustering methods are robust to noise and outliers, and can reduce the dimensionality of the data, they typically rely on domain knowledge to manually construct high-quality features. Sequence to sequence (seq2seq) models can learn representations from sequence data in an unsupervised manner by designing appropriate learning objectives, such as reconstruction and context prediction. When applying seq2seq to time series clustering, obtaining a representation that effectively represents the temporal dynamics of the sequence, multi-scale features, and good clustering properties remains a challenge.
Belief Evolution Network: Probability Transformation of Basic Belief Assignment and Fusion Conflict Probability
Zhou, Qianli, Huang, Yusheng, Deng, Yong
We give a new interpretation of basic belief assignment transformation into probability distribution, and use directed acyclic network called belief evolution network to describe the causality between the focal elements of a BBA. On this basis, a new probability transformations method called full causality probability transformation is proposed, and this method is superior to all previous method after verification from the process and the result. In addition, using this method combined with disjunctive combination rule, we propose a new probabilistic combination rule called disjunctive transformation combination rule. It has an excellent ability to merge conflicts and an interesting pseudo-Matthew effect, which offer a new idea to information fusion besides the combination rule of Dempster.
Learning Representations for Time Series Clustering
Ma, Qianli, Zheng, Jiawei, Li, Sen, Cottrell, Gary W.
Time series clustering is an essential unsupervised technique in cases when category information is not available. It has been widely applied to genome data, anomaly detection, and in general, in any domain where pattern detection is important. Although feature-based time series clustering methods are robust to noise and outliers, and can reduce the dimensionality of the data, they typically rely on domain knowledge to manually construct high-quality features. Sequence to sequence (seq2seq) models can learn representations from sequence data in an unsupervised manner by designing appropriate learning objectives, such as reconstruction and context prediction. When applying seq2seq to time series clustering, obtaining a representation that effectively represents the temporal dynamics of the sequence, multi-scale features, and good clustering properties remains a challenge.