<div dir="auto"><p style="margin:0px 0px 1em;padding:0px;border:0px;font-size:15px;line-height:inherit;font-family:arial,"helvetica neue",helvetica,sans-serif;vertical-align:baseline;color:rgb(36,39,41)">The documentation says:</p><blockquote style="margin:0px 0px 10px;padding:10px;border-width:0px 0px 0px 2px;border-left-style:solid;border-left-color:rgb(255,235,142);font-size:15px;line-height:inherit;font-family:arial,"helvetica neue",helvetica,sans-serif;vertical-align:baseline;quotes:none;background-color:rgb(255,249,227);color:rgb(36,39,41)"><p style="margin:0px;padding:0px;border:0px;font-style:inherit;font-variant:inherit;font-weight:inherit;font-size:inherit;line-height:inherit;font-family:inherit;vertical-align:baseline">The loss function to be used. Defaults to ‘hinge’, which gives a linear SVM. The ‘log’ loss gives logistic regression, a probabilistic classifier. ‘modified_huber’ is another smooth loss that brings tolerance to outliers as well as probability estimates.</p></blockquote><p style="margin:0px 0px 1em;padding:0px;border:0px;font-size:15px;line-height:inherit;font-family:arial,"helvetica neue",helvetica,sans-serif;vertical-align:baseline;color:rgb(36,39,41)">When we use 'modified_huber' loss function, which classification algorithm is used? Is it SVM? If yes, how come it is able to give probability estimates, which is something it can't do with hinge loss?</p><p style="margin:0px 0px 1em;padding:0px;border:0px;font-size:15px;line-height:inherit;font-family:arial,"helvetica neue",helvetica,sans-serif;vertical-align:baseline;color:rgb(36,39,41)">Regards,</p><p style="margin:0px 0px 1em;padding:0px;border:0px;font-size:15px;line-height:inherit;font-family:arial,"helvetica neue",helvetica,sans-serif;vertical-align:baseline;color:rgb(36,39,41)">Vikas</p></div>