Machine Learning Techniques for Time Series Classification

Beschreibung

https://cuvillier.de/de/shop/publications/8834-machine-learning-techniques-for-time-series-classification Classification of time series is an important task in various fields, e.g., medicine, finance, and industrial applications. This work discusses strong temporal classification using machine learning techniques. Here, two problems must be solved: the detection of those time instances when the class labels change and the correct assignment of the labels. For this purpose the scenario-based random forest algorithm and a segment and label approach are introduced. The latter is realized with either the augmented dynamic time warping similarity measure or with interpretable generalized radial basis function classifiers. The main application presented in this work is the detection and categorization of car crashes using machine learning. Depending on the crash severity different safety systems, e.g., belt tensioners or airbags must be deployed at time instances when the best-possible protection of passengers is assured.

Autor*in

Michael Botsch

Themen in »Machine Learning Techniques for Time Series Classification«

time series car crashes machine learning airbags protection of passengers

Stimmen zu »Machine Learning Techniques for Time Series Classification«

Details

Link teilen