1 March 2021 in IEEE Transactions on Fuzzy Systems
The existing long-term time-series forecasting methods based on the neural networks suffer from multiple limitations, such as accumulated errors and diminishing temporal correlation, which compromise the prediction quality. To overcome these shortcomings, in this article, we build trend fuzzy granulation-based long short-term memory (LSTM) neural networks to carry out long-term forecasting, where data points with consistent trend characteristics, including trend change, fluctuation range, and trend persistence, are predicted in unison rather than individually. Noticing that these trend characteristics are more urgently needed than just magnitude information, a question on how to granulate a time series into a granular time series consists of meaningful granules containing trend information that comes to be a crucial step. Only if the established granules fit the varying patterns of time series at utmost, such a granulation can make sense for the improvement of the forecasting accuracy; thus, an optimization method of trend-oriented fuzzy granulation is proposed to meliorate the granulation results. With the built trend fuzzy granulation-based LSTM networks, the successive iterations of one-step forecasting are prevented, and the prediction errors of data within a granule will not further increase. This is the first attempt to build trend fuzzy granule-based LSTM to predict the trend characteristics. Experiments on publicly available time series show good performance of the proposed model.
23 October 2023 in IEEE Transactions on Geoscience and Remote Sensing
The multiple change detection (MCD) of hyperspectral images (HSIs) is the process of detecting change areas and providing “from–to” change information of HSIs obtained from the same area at different times. HSIs have hundreds of spectral bands and contain a large amount of spectral information. However, current deep-learning-based MCD methods do not pay special attention to the interspectral dependency and the effective spectral bands of various land covers, which limits the improvement of HSIs’ change detection (CD) performance. To address the above problems, we propose a spectrum-induced transformer-based feature learning (STFL) method for HSIs. The STFL method includes a spectrum-induced transformer-based feature extraction module (STFEM) and an attention-based detection module (ADM). First, the 3D-2D convolutional neural networks (CNNs) are used to extract deep features, and the transformer encoder (TE) is used to calculate self-attention matrices along the spectral dimension in STFEM. Then, the extracted deep features and the learned self-attention matrices are dot-multiplied to generate more discriminative features that take the long-range dependency of the spectrum into account. Finally, ADM mines the effective spectral bands of the difference features learned from STFEM by the attention block (AB) to explore the discrepancy of difference features and uses the softmax function to identify multiple changes. The proposed STFL method is validated on two hyperspectral datasets, and their experiments illustrate the superiority of the proposed STFL method over the currently existing MCD methods.
30 December 2012 in International Journal of Machine Learning and Cybernetics