SLIDE 25 Exploratory Behaviour in CHS
- [Cartright et al., 2011] argue
that a portion of health-directed searches are exploratory in
- nature. These could be divided
into two iterative phases
- evidence-directed: findings
are fused to construct a list of potential explanatory diagnoses ranked by likelihood
- hypothesis-directed: list of
diagnoses used to guide collection of additional evidence, to validate/choose hypotheses.
–
Hypothesis- Directed Inference Evidence-Directed Inference Stop?
SAT/DSAT, Action Diagnostic Intent Informational Intent Yes No
Stop?
Yes No Initial intention (diagnosis, information) Initial intention (diagnosis, information) SAT/DSAT, Action
SIGIR’11 –
q1 q2 q3 q4
Frames: Actions:
Symptoms: [headache,0] Causes: [stress,0], [concussion,1] Remedies: None Symptoms: [headache,1] Causes: [stress,1], [concussion,2] Remedies: [aspirin,0]
[stress headache] [concussion] [aspirin]
...
symptom “back pain” and rem dy “exercise.” We define a user‟s focus of attention over a single action. Each frame consists We see large variations in users‟ search behaviors, including how
- terms/phrases such as “ache” and “dizziness”, and;
“pain in” or “causes of”
- terms/phrases such as “acid reflux” and “sinusitis”, and;
“symptoms of” or “diagnosis of”
- terms such as “treatment”, “clinic”, and “doctor”, and;
“cure for” or “treatment for”
Images from [Cartright et al., 2011]
