Using Data Analysis to Study ME/CFS and Long Covid

[Themos Kalafatis]

Dear All,

cc @Pyrrhus @Learner1 @Martin aka paused||M.E. @SlamDancin

As a patient and now a researcher, I am grateful to be a recipient of the 2021 Stupski Awards from SolveCFS. I will be giving an interview to Leslie Phillips on Friday , January 21st at 10:00am PT regarding the analysis i will be performing to ME/CFS and LongCOVID patient data using Machine Learning and Natural Language Processing methods to identify factors that differentiate symptom severity and also using cutting-edge temporal agortihms that identify patterns of symptoms across time.

To be discussed :

-How i used Machine Learning in 2011 to analyze the severity of my symptoms in data that i logged for a total of 434 days and how these patterns helped me unserstand more my symptoms.

-How temporal analysis identified the "vicious circle" of my symptoms and why a specific symptom deserved particular attention.

-How Natural Language Processing can help us build a high-quality patient dataset for further analysis

-Why it is important for ME/CFS and LongCOVID patients to share their story with as many details as possible using text notes.

-What kind of knowledge we expect to have after analyzing patient data

You may register for the event here :

https://solvecfs.org/event/using-data-analysis-to-study-me-cfs-and-long-covid-session-1/


Disclaimer : I am a patent owner of a methodology that uses Artificial Intelligence and Text Analysis methods to research any medical syndrome, disease or group of symptoms

https://patents.justia.com/patent/10748663

 

[Pyrrhus]

As a patient and now a researcher, I am grateful to be a recipient of the 2021 Stupski Awards from SolveCFS.

Congratulations!

 

[Martin aka paused||M.E.]

Dear All,

cc @Pyrrhus @Learner1 @Martin aka paused||M.E. @SlamDancin

As a patient and now a researcher, I am grateful to be a recipient of the 2021 Stupski Awards from SolveCFS. I will be giving an interview to Leslie Phillips on Friday , January 21st at 10:00am PT regarding the analysis i will be performing to ME/CFS and LongCOVID patient data using Machine Learning and Natural Language Processing methods to identify factors that differentiate symptom severity and also using cutting-edge temporal agortihms that identify patterns of symptoms across time.

To be discussed :

-How i used Machine Learning in 2011 to analyze the severity of my symptoms in data that i logged for a total of 434 days and how these patterns helped me unserstand more my symptoms.

-How temporal analysis identified the "vicious circle" of my symptoms and why a specific symptom deserved particular attention.

-How Natural Language Processing can help us build a high-quality patient dataset for further analysis

-Why it is important for ME/CFS and LongCOVID patients to share their story with as many details as possible using text notes.

-What kind of knowledge we expect to have after analyzing patient data

You may register for the event here :

https://solvecfs.org/event/using-data-analysis-to-study-me-cfs-and-long-covid-session-1/


Disclaimer : I am a patent owner of a methodology that uses Artificial Intelligence and Text Analysis methods to research any medical syndrome, disease or group of symptoms

https://patents.justia.com/patent/10748663

That’s very interesting. Maybe you are interested in a zoom call with me?

 

[Themos Kalafatis]

Thank you @Martin aka paused||M.E. , i will PM you very soon

 

[Learner1]

Thank you Themos. I am PMing you a couple questions 😉

 

[anne_likes_red]

Dear All,

cc @Pyrrhus @Learner1 @Martin aka paused||M.E. @SlamDancin

As a patient and now a researcher, I am grateful to be a recipient of the 2021 Stupski Awards from SolveCFS. I will be giving an interview to Leslie Phillips on Friday , January 21st at 10:00am PT regarding the analysis i will be performing to ME/CFS and LongCOVID patient data using Machine Learning and Natural Language Processing methods to identify factors that differentiate symptom severity and also using cutting-edge temporal agortihms that identify patterns of symptoms across time.

To be discussed :

-How i used Machine Learning in 2011 to analyze the severity of my symptoms in data that i logged for a total of 434 days and how these patterns helped me unserstand more my symptoms.

-How temporal analysis identified the "vicious circle" of my symptoms and why a specific symptom deserved particular attention.

-How Natural Language Processing can help us build a high-quality patient dataset for further analysis

-Why it is important for ME/CFS and LongCOVID patients to share their story with as many details as possible using text notes.

-What kind of knowledge we expect to have after analyzing patient data

You may register for the event here :

https://solvecfs.org/event/using-data-analysis-to-study-me-cfs-and-long-covid-session-1/


Disclaimer : I am a patent owner of a methodology that uses Artificial Intelligence and Text Analysis methods to research any medical syndrome, disease or group of symptoms

https://patents.justia.com/patent/10748663

Congrats!!
I've been a big fan of what you're doing a long time now.
Missed the webinar unfortunately. Let us know if it was recorded. Anne.

 

[Consul]

@Themos Kalafatis very interresting work you are doing. So got any timeline for this?

And what do you think the things you did to wrestle out of me/cfs suggets about the underlying cause of the disease? (if anything)

 

[Themos Kalafatis]

@Consul Thank you for your supportive comment.

As discussed, I am currently in the process of enriching the dataset. The work that a Machine Learning algorithm has to do becomes much easier when the input data are enriched in particular ways, it enables us to analyze data from different "perspectives". Also, i am expecting more data to come in.

I am very much a fan of a personalized approach using genomic data and the AI system i am using suggests a mixture of immuno+metabolic+mitochondrial issues. I have a number of hypotheses that i am hoping that some researchers may look at.

The truth is that i do not know why whatever i did worked for me. I believe that i simply made -say- 70% of the right steps which helped me overcome my problems in the long run. I was very lucky that whatever intervention i was doing had an immediate positive / negative effect (if it had one). Logging of my symptoms also led me to get to know my "body signals" extremely well.

Some -hopefully- interesting info :

- It took me many months to finally accept that i was able to live a normal life.

- The photo i use in my profile is from a hike to Drakolimni (https://en.wikipedia.org/wiki/Drakolimni) i did which required several hours of walking up a mountain. The fact that i did not crash after so many hours of walking (approx. 5 hours) had me finally convinced that i could live a normal life.

-I still consider myself a patient since if i do the wrong things (for me) , i will crash.

-I would love to be able to perform "Personal Data Mining" in every patient who submitted data, meaning to extract symptom patterns on a patient level.