Research Article
HPM: A Hybrid Model for User’s Behavior Prediction Based on N-Gram Parsing and Access Logs
| Input: N-gram associated click-graph (NC-graph) | | Output: weighted N-grams corresponding to distinct URLs stored in matrix WL | | Begin | (1) | Create a matrix WL of order m × n//m ⟶ no. of distinct N-grams of all the queries of PL and n ⟶ no. of URLs of PL | (2) | ;//elements of WL | (3) | For each URL in NC-graph do | (4) | //weight of N-gram associated with query q corresponding to URL u | (5) | For each N-gram in NC-graph do | (6) | Cn,u = Cq,u; // | (7) | ; | (8) | End For | (9) | For each N-gram in NC-graph do | (10) | and //normalization of calculated weights | (11) | Store in WL; | (12) | EndFor | (13) | EndFor | (14) | Return WL; | (15) | End |
|