Predictivity of Category Based Human Navigation and the Effect of Navigation Path Length on the Prediction Accuracy in Knowledge Networks

In web sites containing large amount of information, systems which help visitors access the contents they are interested in, significantly increase the usability of the sites. Knowing the category of the content that the user wants to access is important to offer more qualified suggestions and advertisements to the user. On the other hand, as the user continues to navigate in the site, more navigation data is obtained from him, which may potentially be used for better predictions.In this study, the predictability of users' category-based navigation and the effect of the length of the navigation path on the prediction accuracy in knowledge networks is investigated using LSTM classifiers.