Display similarities with percetange

virank · 2024-08-11T13:37:49.449Z

I am sorry for always asking, I don’t want to seem like a disturbance. I am still a newbie.
Well, i am in a dilema where i am looking for the correct way of doing something

I want to do display on repeating group a source of data date are very similar to a specific data so that when doing researches I can refer to other done before the onde I am doing and diplay the percentage of similarities.

My current app, is a project for patient record management
I want to be able while saving different type of data during my consultation to be able to see if there are similar cases that had happened before so that I can i can refer myself to it and gain an insight on how I could handle the present case. I hope I make sense. But I dont want to stop there, I want to be able to display at how many percentages similarities are.
This is my current table configuration to where I want similarities to be drawn from

Patient name
Plainte
Histoire de la maladie
CA
Clinique
Diagnostique

I want to be able to quantify how many pecetage of similarities are found on each cases. And display them according to their percentage from higher to lower.
I hope I find someone to help

senecadatabase · 2024-08-11T14:50:46.772Z

You’re trying to get percentages based on text.

That is more complicated, but can be done.

However, the comparison formula would involve Levenshtein distance, Cosine similarity, and Jaccard similarity.

Are you familiar with any of those?

Also, I would be very careful about using Bubble for patient information since it’s not HIPPA compliant. Just a thought.

Maybe someone with more experience in Bubble has a better idea.

virank · 2024-08-11T15:02:52.446Z

Thank you @senecadatabase ,
Well since it is just a project and is not based in the USA I do not have to worry about HIPAA. Thanks for the tip.

As stated earlier, my aim is to compare my current data to a list of datas in a repeating group so I can see how similar they. SOrt of looking for plagiarism and how many percent that plagiarism has been done. The main reason of quest is to be able to display a list in a repeating group of patient record that are similar to the current one I could be working on and the level of similarities which can be in percentage. I checked many plugins sort of offering something similar but many working on the comparison of two specific text. What I need I to do a search my entire database and extract those that look similar.
No I am not familiar with Levenhstein distance more Cosine similarity.
If you could help me maybe explaining, a possibility, I will be very grateful

senecadatabase · 2024-08-11T15:13:57.813Z

Ok, sorry, I was basing my answer about HIPPA in the U.S.

If it’s ok in your country I guess that doesn’t apply.

Myself, I think the easiest approach would be to assign numerical values to the items.

You could use option sets.

Once you’ve assigned numerical values, it would be easier to calculate percentages.

Basing percentages on text data involves breaking the text down and calculating it which is much more detailed.

virank · 2024-08-11T15:16:36.637Z

Thank you dear
Well how do I do that

senecadatabase · 2024-08-11T15:27:58.400Z

Things like if it’s one illness would be a 1.

Another illness would be a 2.

One clinic would be a 1 etc.

Once you assign a numerical value you can do calculations easier.

virank · 2024-08-11T15:28:39.976Z

Thank you dear

senecadatabase · 2024-08-11T15:31:33.218Z

Hope you do good with your project.

If you need anymore help, just ask.

Maybe someone else with more experience with Bubble has a better answer.

system · 2024-10-20T13:38:10.128Z

This topic was automatically closed after 70 days. New replies are no longer allowed.