Implications of Computational Cognitive Models for Information Retrieval

and determine from that abstract text that they are not interested in reading the full text. A request for the full text is not necessarily a relevance judgment, but is an indication that the user wanted to read more of the document than just the abstract. For all of the approaches, the documents that the HAM-TMC users viewed were assigned a ranking score by one of the models in the experiments. The ranking scores were used to determine how many of the extracted pairwise judgments from the document download pairwise judgment data set or document click pairwise judgment data set were correctly ordered (e.g., documentA is preferred over documentB) based on the ranking scores. Figure 5.7. Experiment for computing information scent Query issued by user All documents viewed by user in response to query