Chain-structured neural architecture search for financial time series forecasting
By Denis Levchenko, Efstratios Rappos, Shabnam Ataee, Biagio Nigro, Stephan Robert
Rating
1084
Battle Count: 84
Relevance
8/10
Directly applicable to financial time series prediction, though performance improvements are modest
Implementation Complexity
7/10
Requires understanding of multiple NAS strategies and deep learning architectures
Reproducibility
3/5
Code not provided but methods described in detail. Datasets are proprietary.
About this paper
Methodology: Chain-structured neural architecture search. Problem types: Time Series Forecasting, Binary Classification.
The interactive Everscope explorer (charts, battles, favorites) loads below.