These preprocessing tasks resemble those in any other web usage mining problem and are discussed in detail in Hussain, Asghar, and Masood (2010).The original server logs are cleansed, formatted, and then grouped into meaningful sessions before use in web usage mining. A session can be described as the self-care activities performed by an elder between the start and the end of the ComCare session. Therefore, session identification is the process of segmenting the access log of each elder into individual access sessions. The activity stay time-based method developed by Liu and Kešelj (2007) was applied for session identification in this study (Liu & Kešelj, 2007).This method limits the time spent on a function of ComCare to a specified threshold. If the time between the request most recently assigned to a session and the next request
from the elder exceeds the threshold, a new access session
is assumed. A 10 min activity-stay time is considered a conservative
threshold for capturing the time for loading and studying page
content (Liu & Kešelj, 2007). However, this study uses the average
use time of the function of all sessions as a threshold for function
use time.