When you assign a weight to the child of a topic which uses a concept operator, you specify the relative contribution of that child to the overall score produced by a topic. The higher the weight you assign to the child, the higher selected documents which contain that child will appear in the list of results. Thus, weights directly affect the importance, or ranking, of selected documents.
For example, assume you have the following topic:
The evidence topics 80286 and 80386 (which describe the microprocessors used in PC products) have an automatic weight assignment of 1.00. The evidence topics 486, 386, and 286 have a relatively high probability of referring to their parent topic, so these evidence topics are assigned weights of 0.80. The evidence topic clone may or may not refer to PC clones at all; therefore, this evidence topic is assigned a weight of 0.40.
A search agent using this topic and its assigned weights might produce the following scores for the matched documents:
If you change the weights of each evidence topic, the importance of your selection results are affected, as well. In this example, if you change the weights of the evidence topic 486 to 0.60, the evidence topic 386 to 0.45, the evidence topic 286 to 0.35, and the evidence topic clone to 0.20, your selected document scores will change as follows:
