Zezhong Zhang

A server accesses a plurality of users' sessions with the web server. Each user session indicating a page flow of a corresponding user session for a plurality of web pages provided by the web server. The server generates a learning model using a neural network based on the plurality of users' sessions. The learning model is configured to predict a next user activity based on a current page flow of a current user session. The next user activity indicating one of continuing the current user session by visiting another web page provided by the web server and ending the current user session. The server dynamically adjusts a content of a web page based on the predicted next user activity.

A server accesses a plurality of users' sessions with the web server. Each user session indicating a page flow of a corresponding user session for a plurality of web pages provided by the web server. The server generates a learning model using a neural network based on the plurality of users' sessions. The learning model is configured to predict a next user activity based on a current page flow of a current user session. The next user activity indicating one of continuing the current user session by visiting another web page provided by the web server and ending the current user session. The server dynamically adjusts a content of a web page based on the predicted next user activity.