@techreport{oai:ipsj.ixsq.nii.ac.jp:00216898, author = {Thanapiol, Phungtua-eng and Yoshitaka, Yamamoto and Thanapiol, Phungtua-eng and Yoshitaka, Yamamoto}, issue = {5}, month = {Mar}, note = {The data stream may contain irrelevant information due to various factors, such as the huge amount of data volume. The technique for removing irrelevant information is called data binning. The data binning is used to sequence the data stream into smaller bins and each of which captures statistical features in the corresponding sub-sequence. The obtained bins are collected into the window, meaning that the number of bins to be collected is determined by the window size. The window size is required to set in advance, whereas the sufficient value is varied with the target data stream to be captured. This paper proposes a novel framework for automatically adjusting the number of bins in the window with a non-parametric metric. We demonstrate our framework to detect unknown transient patterns with the astronomical data stream., The data stream may contain irrelevant information due to various factors, such as the huge amount of data volume. The technique for removing irrelevant information is called data binning. The data binning is used to sequence the data stream into smaller bins and each of which captures statistical features in the corresponding sub-sequence. The obtained bins are collected into the window, meaning that the number of bins to be collected is determined by the window size. The window size is required to set in advance, whereas the sufficient value is varied with the target data stream to be captured. This paper proposes a novel framework for automatically adjusting the number of bins in the window with a non-parametric metric. We demonstrate our framework to detect unknown transient patterns with the astronomical data stream.}, title = {A novel framework of non-parametric for adjusting the window size}, year = {2022} }