MSA: The forgotten index for identifying inappropriate items before computing exploratory item factor analysis

Kaiser’s single-variable measure of sampling adequacy (MSA) is a very useful index for debugging inappropriate items before a factor analysis (FA) solution is fitted to an item-pool dataset for item selection purposes. For reasons discussed in the article, however, MSA is hardly used nowadays in this context. In our view, this is unfortunate. In the present proposal, we first discuss the foundation and rationale of MSA from a ‘modern’ FA view, as well as its usefulness in the item selection process. Second, we embed the index within a robust approach and propose improvements in the preliminary item selection process. Third, we implement the proposal in different statistical programs. Finally, we illustrate its use and advantages with an empirical example in personality measurement.

A Coarse-Grained Regularization Method of Convolutional Kernel for Molten Pool Defect Identification

Machine vision has a wide range of applications in the field of welding. The rise of convolutional neural network (CNN) provides a new way to extract visual features of welding. Due to the limitation of the small size of our molten pool dataset, the regularization of the CNN model is necessary to prevent overfitting. We propose a coarse-grained regularization method for convolution kernels (CGRCKs), which is designed to maximize the difference between convolution kernels in the same layer. The algorithm performance was tested on our self-made dataset and other public datasets. The results show that the CGRCK method can extract multi-faceted features. Compared with L1 or L2 regularization, the proposed method works great on CNNs and introduces little overhead cost to the training.