Improving ELM-Based Time Series Classification by Diversified Shapelets Selection Qifa Sun 1 , Qiuyan Yan 1,2(&) , Xinming Yan 1 , Wei Chen 1 , and Wenxiang Li 3 1 School of Computer Science and Technology, China University of Mining Technology, Xuzhou 221116, China sunqifa@live.com, {yanqy,yanxm,chenw}@cumt.edu.cn 2 School of Safty Engineering, China University of Mining Technology, Xuzhou 221116, China 3 School of Information Science and Engineering, Wuhan University of Science and Technology, Wuhan 430081, China liwx2006@hotmail.com Abstract. ELM is an ef ficient neural network which has extremely fast learning capacity and good generalization capability. However, ELM fails to measure up the task of time series classification because it hard to extract the features and characters of time series data. Especially, many time series has trend features which cannot be abstracted by ELM thus lead to accuracy decreasing. Although through selection good features can improve the interpretability and accuracy of ELM, canonical methods either fails to select the most representative and interpretative features, or determine the number of features parameterized. In this paper, we propose a novel method by selection diversified top-k shapelets to improve the interpretability and accuracy of ELM. There are three contributions of this paper: First, we put forward a trend feature symbolization method to extract the trend information of time series; Second, the trend feature symbolic expressions are mapped into a shapelet can- didates set and a diversified top-k shapelets selection method, named as Div- TopkShapelets, are proposed to find the most k distinguish shapelets; Last, we proposed an iterate ELM method, named as DivShapELM, automatically deter- mining the best shapelets number and getting the optimum ELM classifier. The experimental results show that our proposed methods significantly improves the effectiveness and interpretability of ELM. Keywords: Extreme Learning Machine  Time series classification  Shapelets  Diversified query 1 Introduction Extreme Learning Machine (ELM for short), based on single-hidden layer feedforward neural networks (SLFNs), was proposed for addressing the slow speed of traditional neural networks. ELM has the extremely fast learning capability and good general- ization capability through assign the weights connecting inputs to hidden nodes ran- domly. The weights between hidden nodes and outputs are learned in a single step, © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2017 J.-H. Lee and S. Pack (Eds.): QShine 2016, LNICST 199, pp. 446–456, 2017. DOI: 10.1007/978-3-319-60717-7_44