Detach-ROCKET: Sequential feature selection for time series classification with random convolutional kernels
Authors:
Gonzalo Uribarri,
Federico Barone,
Alessio Ansuini,
Erik Fransén
Abstract:
Time Series Classification (TSC) is essential in fields like medicine, environmental science, and finance, enabling tasks such as disease diagnosis, anomaly detection, and stock price analysis. While machine learning models like Recurrent Neural Networks and InceptionTime are successful in numerous applications, they can face scalability issues due to computational requirements. Recently, ROCKET h…
▽ More
Time Series Classification (TSC) is essential in fields like medicine, environmental science, and finance, enabling tasks such as disease diagnosis, anomaly detection, and stock price analysis. While machine learning models like Recurrent Neural Networks and InceptionTime are successful in numerous applications, they can face scalability issues due to computational requirements. Recently, ROCKET has emerged as an efficient alternative, achieving state-of-the-art performance and simplifying training by utilizing a large number of randomly generated features from the time series data. However, many of these features are redundant or non-informative, increasing computational load and compromising generalization. Here we introduce Sequential Feature Detachment (SFD) to identify and prune non-essential features in ROCKET-based models, such as ROCKET, MiniRocket, and MultiRocket. SFD estimates feature importance using model coefficients and can handle large feature sets without complex hyperparameter tuning. Testing on the UCR archive shows that SFD can produce models with better test accuracy using only 10\% of the original features. We named these pruned models Detach-ROCKET. We also present an end-to-end procedure for determining an optimal balance between the number of features and model accuracy. On the largest binary UCR dataset, Detach-ROCKET improves test accuracy by 0.6\% while reducing features by 98.9\%. By enabling a significant reduction in model size without sacrificing accuracy, our methodology improves computational efficiency and contributes to model interpretability. We believe that Detach-ROCKET will be a valuable tool for researchers and practitioners working with time series data, who can find a user-friendly implementation of the model at \url{https://github.com/gon-uri/detach_rocket}.
△ Less
Submitted 24 June, 2024; v1 submitted 25 September, 2023;
originally announced September 2023.
A novel multi-layer modular approach for real-time fuzzy-identification of gravitational-wave signals
Authors:
Francesco Pio Barone,
Daniele Dell'Aquila,
Marco Russo
Abstract:
Advanced LIGO and Advanced Virgo ground-based interferometers are instruments capable to detect gravitational wave signals exploiting advanced laser interferometry techniques. The underlying data analysis task consists in identifying specific patterns in noisy timeseries, but it is made extremely complex by the incredibly small amplitude of the target signals. In this scenario, the development of…
▽ More
Advanced LIGO and Advanced Virgo ground-based interferometers are instruments capable to detect gravitational wave signals exploiting advanced laser interferometry techniques. The underlying data analysis task consists in identifying specific patterns in noisy timeseries, but it is made extremely complex by the incredibly small amplitude of the target signals. In this scenario, the development of effective gravitational wave detection algorithms is crucial. We propose a novel layered framework for real-time detection of gravitational waves inspired by speech processing techniques and, in the present implementation, based on a state-of-the-art machine learning approach involving a hybridization of genetic programming and neural networks. The key aspects of the newly proposed framework are: the well structured, layered approach, and the low computational complexity. The paper describes the basic concepts of the framework and the derivation of the first three layers. Even if the layers are based on models derived using a machine learning approach, the proposed layered structure has a universal nature. Compared to more complex approaches, such as convolutional neural networks, which comprise a parameter set of several tens of MB and were tested exclusively for fixed length data samples, our framework has lower accuracy (e.g., it identifies 45% of low signal-to-noise-ration gravitational wave signals, against 65% of the state-of-the-art, at a false alarm probability of $10^{-2}$), but has a much lower computational complexity and a higher degree of modularity. Furthermore, the exploitation of short-term features makes the results of the new framework virtually independent against time-position of gravitational wave signals, simplifying its future exploitation in real-time multi-layer pipelines for gravitational-wave detection with new generation interferometers.
△ Less
Submitted 16 December, 2023; v1 submitted 13 June, 2022;
originally announced June 2022.
