
This paper proposes efficient algorithms to prune redundant random convolution kernels in time series classification models like ROCKET and MINIROCKET, without significantly degrading their performance.


coremsg

efficient-pruning-of-random-convolution-kernels-for-time-series-classification-using-a-feature-selection-approach


Efficient Pruning of Random Convolution Kernels for Time Series Classification using a Feature Selection Approach
