For the Filter Bubble project, we have browsing history data from Chrome and from Safari. Users who give their browsing history to us are asked to sync all their devices before the upload so that the data contains their history from both desktop and mobile devices.
Chrome and from Safari store slightly different information about URLs. In particular, the Safari browser history database contains the column 'origin' which is a number related to the device used to access that URL. Chrome does not store this information.
Is it possible to build a model to predict the origin of a Chrome URL (i.e. mobile or desktop) based on the URL (and, potentially, other data such a timestamp) when it is trained on the labelled Safari data?
For the Filter Bubble project, we have browsing history data from Chrome and from Safari. Users who give their browsing history to us are asked to sync all their devices before the upload so that the data contains their history from both desktop and mobile devices.
Chrome and from Safari store slightly different information about URLs. In particular, the Safari browser history database contains the column 'origin' which is a number related to the device used to access that URL. Chrome does not store this information.
Is it possible to build a model to predict the origin of a Chrome URL (i.e. mobile or desktop) based on the URL (and, potentially, other data such a timestamp) when it is trained on the labelled Safari data?