Unsupervised feature selection for multi-view data with missing views is a challenging task. Most existing approaches handle this by first imputing the missing data and then performing feature ...