A Sensitivity Analysis of (and Practitioners' Guide to) Convolutional\n Neural Networks for Sentence Classification

Convolutional Neural Networks (CNNs) have recently achieved remarkably strong\nperformance on the practically important task of sentence classification (kim\n2014, kalchbrenner 2014, johnson 2014). However, these models require\npractitioners to specify an exact model architecture and set accompanying\nhyperparameters, including the filter region size, regularization parameters,\nand so on. It is currently unknown how sensitive model performance is to\nchanges in these configurations for the task of sentence classification. We\nthus conduct a sensitivity analysis of one-layer CNNs to explore the effect of\narchitecture components on model performance; our aim is to distinguish between\nimportant and…