Abstract:Recent years have witnessed the unprecedented rising of time series from almost all kindes of academic and industrial fields. Various types of deep neural network models have been introduced to time series analysis, but the important frequency information is yet lack of effective modeling. In light of this, in this paper we propose a wavelet-based neural network structure called multilevel Wavelet Decomposition Network (mWDN) for building frequency-aware deep learning models for time series analysis. mWDN preserves the advantage of multilevel discrete wavelet decomposition in frequency learning while enables the fine-tuning of all parameters under a deep neural network framework. Based on mWDN, we further propose two deep learning models called Residual Classification Flow (RCF) and multi-frequecy Long Short-Term Memory (mLSTM) for time series classification and forecasting, respectively. The two models take all or partial mWDN decomposed sub-series in different frequencies as input, and resort to the back propagation algorithm to learn all the parameters globally, which enables seamless embedding of wavelet-based frequency analysis into deep learning frameworks. Extensive experiments on 40 UCR datasets and a real-world user volume dataset demonstrate the excellent performance of our time series models based on mWDN. In particular, we propose an importance analysis method to mWDN based models, which successfully identifies those time-series elements and mWDN layers that are crucially important to time series analysis. This indeed indicates the interpretability advantage of mWDN, and can be viewed as an indepth exploration to interpretable deep learning.

Fully learnable deep wavelet transform for unsupervised monitoring of high-frequency time series

MultiWave: Multiresolution Deep Architectures through Wavelet Decomposition for Multivariate Time Series Prediction

WEITS: A Wavelet-enhanced residual framework for interpretable time series forecasting

Weighted Nonlinear Dynamic System for Deep Extraction of Nonlinear Dynamic Latent Variables and Industrial Application

Wavelet-Driven Spatiotemporal Predictive Learning: Bridging Frequency and Time Variations

WaveForM: Graph Enhanced Wavelet Learning for Long Sequence Forecasting of Multivariate Time Series

W-FENet: Wavelet-based Fourier-Enhanced Network Model Decomposition for Multivariate Long-Term Time-Series Forecasting

Representing the dynamics of high-dimensional data with non-redundant wavelets

Multilevel Wavelet Decomposition Network for Interpretable Time Series Analysis

A New Deep Learning Framework for HF Signal Detection in Wideband Spectrogram

Wavelet gated multiformer for groundwater time series forecasting

Interpretable Real-Time Monitoring of Pipeline Weld Crack Leakage Based on Wavelet Multi-Kernel Network

Deep Learning in the Wavelet Domain

Wave2Vec: Learning Deep Representations for Biosignals

T-WaveNet: Tree-Structured Wavelet Neural Network for Sensor-Based Time Series Analysis

Automatic Waveform-Based Source-Location Imaging Using Deep Learning Extracted Microseismic Signals

Robust Time Series Denoising with Learnable Wavelet Packet Transform

Wavelet Integrated CNN With Dynamic Frequency Aggregation for High-Speed Train Wheel Wear Prediction

Dynamic detection of anomalous regions within distributed acoustic sensing data streams using locally stationary wavelet time series

Wave2Vec: Deep Representation Learning for Clinical Temporal Data

An Unsupervised Deep Learning Approach for the Wave Equation Inverse Problem