"New login location detected":"New login location detected",
"A new login into your account was detected. The IP address %s was classified as suspicious. If this was you, you can ignore this message. Otherwise you should change your password.":"A new login into your account was detected. The IP address %s was classified as suspicious. If this was you, you can ignore this message. Otherwise you should change your password.",
"You can get more info by pressing the button which will open %s and show info about the suspicious IP-address.":"You can get more info by pressing the button which will open %s and show info about the suspicious IP-address.",
"The suspicious login app is enabled on this instance. It will keep track of IP addresses users successfully log in from and build a classifier that warns if a new login comes from a suspicious IP address.":"The suspicious login app is enabled on this instance. It will keep track of IP addresses users successfully log in from and build a classifier that warns if a new login comes from a suspicious IP address.",
"Training data statistics":"Training data statistics",
"So far the app has captured {total} logins (including client connections), of which {distinct} are distinct (IP, UID) tuples.":"So far the app has captured {total} logins (including client connections), of which {distinct} are distinct (IP, UID) tuples.",
"Classifier model statistics":"Classifier model statistics",
"No classifier model has been trained yet. This most likely means that you just enabled the app recently. Because the training of a model requires good data, the app waits until logins of at least {days} days have been captured.":"No classifier model has been trained yet. This most likely means that you just enabled the app recently. Because the training of a model requires good data, the app waits until logins of at least {days} days have been captured.",
"During evaluation, the latest model (trained {time}) has shown to capture {recall}% of all suspicious logins (recall), whereas {precision}% of the logins classified as suspicious are indeed suspicious (precision). Below you see a visualization of historic model performance.":"During evaluation, the latest model (trained {time}) has shown to capture {recall}% of all suspicious logins (recall), whereas {precision}% of the logins classified as suspicious are indeed suspicious (precision). Below you see a visualization of historic model performance.",