Research Article
Text Matching and Categorization: Mining Implicit Semantic Knowledge from Tree-Shape Structures
Input: text | Output: SemGraph | () scan sentence | () IF sentence is not empty THEN | () create | () ELSE exist | () FOREACH IN () DO | () IF = NIL THEN | () add to | () Count() = 1 | () ElSE | () IF does not appear in THEN | () add to | () add all edges of to | () Count() = 1 | () IF appears in THEN | () Count() = Count() + 1 | () FOREACH two linked nodes IN DO | () IF the Counts of them are equal THEN | () set the direction of the pointer randomly | () IF the Counts of them are unequal THEN | () set the direction of the pointer from the node with a bigger Count to the one with a smaller Count |
|