The field of time series forecasting and anomaly detection is rapidly advancing with the development of new models and techniques. Recent research has focused on improving the accuracy and efficiency of forecasting models, particularly in multivariate time series data. The use of transformer-based models has shown promising results, with methods such as delegate token attention and physics-informed attention mechanisms achieving state-of-the-art performance. Additionally, the incorporation of structural similarity and multi-scale feature extraction has improved the detection of anomalies in time series data. Noteworthy papers include FRAUDGUESS, which proposes a novel approach to detecting and explaining new types of fraud in financial data, and Pi-Transformer, which presents a physics-informed transformer for time series anomaly detection. Other notable papers include AdaMixT, which introduces a novel architecture for multivariate time series forecasting, and StrAD, which proposes a structure-enhanced anomaly detection approach.
Advances in Time Series Forecasting and Anomaly Detection
Sources
FRAUDGUESS: Spotting and Explaining New Types of Fraud in Million-Scale Financial Data
Tsururu: A Python-based Time Series Forecasting Strategies Library
AdaMixT: Adaptive Weighted Mixture of Multi-Scale Expert Transformers for Time Series Forecasting
Towards Scalable and Structured Spatiotemporal Forecasting
SDGF: Fusing Static and Multi-Scale Dynamic Correlations for Multivariate Time Series Forecasting
A Machine Learning Framework for Pathway-Driven Therapeutic Target Discovery in Metabolic Disorders
GluMind: Multimodal Parallel Attention and Knowledge Retention for Robust Cross-Population Blood Glucose Forecasting
MOMEMTO: Patch-based Memory Gate Model in Time Series Foundation Model
Analyzing the Impact of Credit Card Fraud on Economic Fluctuations of American Households Using an Adaptive Neuro-Fuzzy Inference System
Unsupervised Outlier Detection in Audit Analytics: A Case Study Using USA Spending Data
Transformer Modeling for Both Scalability and Performance in Multivariate Time Series
Pi-Transformer: A Physics-informed Attention Mechanism for Time Series Anomaly Detection
An Improved Time Series Anomaly Detection by Applying Structural Similarity