I have been given the task for my bachelor thesis to investigate how doctors can benefit from the use of negations in sentences that reflect physiological sensor data such as heart- and respiratory rates.
The interest of this mainly comes from the assumption that negations can give more meaning in certain context but also that it may aide doctors in deduction and since it will probably be used in an Ontology that reasons about events.
The problems I am facing right now is how I should go about in generating medical sentences in ECG (Electrocardiography) data. The issues are primarly:
- What they should say.
- How they should say it.
Keep in mind that many factors can influence heart rate, including:
- Activity level
- Fitness level
- Air temperature
- Body position (standing up or lying down, for example)
- Body size
For instance should I compare two sentences like this:
S1: "Patient X have fluctuating heart rates."
S2: "Patient X is not fit and have fluctuating heart rates."
or should I produce sentences similiar to:
S3: "Patient X is obece and have fluctuating heart rates."
S4: "Patient X is not fit and have fluctuating heart rates."
S2 compared to S1 contains more information whilst S3 and S4 contain equal only that S4 is negated, however when dealing with diagnosis, which one would doctors prefer? Maybe I should make the system generate all four?
All input is appreciated!