Context-Aware Computing

(2)Influential Parameters to be Tuned

1) Parameters used in constructing and managing probabilistic context trees

–A weight vector (𝛼1, 𝛼2, 𝛼3, 𝛼4) with ${\displaystyle {\sum }_{i=1}^4{\alpha }_i}=1,$ used in computing the association score of a context activity c with a web page access w in Definition 47. It quantifies the importance of c’s focus duration, focus frequency, temporal distance from the page access, and title similarity between c and w, respectively. It impacts the ranking position of result pages.

–Context focus duration threshold 𝜏c and interleaving window size threshold Δ between c and w in context acquisition. Reducing 𝜏c and enlarging Δ can capture more associated context activities into the probabilistic context tree, and deliver more result ...