RohansModel.extract_sentences
Hello! Right now I am re-writing my model into a clean, easy to use package named RohansModel (I’m taking name suggestions in the comments!). For a user to implement my model, all they have to do is import it from a python file. Then, they can input a custom set of sentences for my model to try and extract from a recording. This is a new feature that I am coding as you read! Next, to extract the sentences, they can call a function from my model called “extract_sentences”, which creates a text file with the start and end times of where the sentences are located in a recording. To make it as simple as possible on the user’s end, it takes a ton of work under the surface.
Today I’m learning how to put all the pieces together, as I move from the testing phase to the production phase. The goal is that when I leave, anyone at the lab can easily load my model, understand what I’ve done, and customize it to their needs. I have a couple hypotheses that could be interesting:
- Does the runtime/accuracy of my model on an audio file directly correlate to the severity of symptoms in that patient?
- If my model shows that a patient is taking longer to say certain sentences, is that a hint of pathological symptoms?
These are questions I may not have time to answer in the short time that’s left, but I think it’s important to ask: how can my model be used in other ways? If it’s customizable and easy to use, the other lab researchers can find out!