Thyroid Detection

A simple ML based website which predicts the thyroid type out of the 4 classes - “negative”, “compensated hypothyroid”, “primary hypothyroid”, “secondary hypothyroid” using various thyroid tests such as ‘TSH’, ‘T3’, ‘TT4’ etc., where negative class means no thyroid related condition.

Application interface screenshot:

Application screenshot showing what the application does:

Features:

1. age - Shows the age of the patient.
2. sex - Shows the gender of the patient.
3. on_thyroxine - Shows whether the patient is on Thyroxine treatment or not.
4. query_on_thyroxine - Shows whether the patient has raised any queries regarding their thyroxine treatment/hormone or not.
5. on_antithyroid_medication - Shows whether the patient is on antithyroid_medication or not.
6. sick - Shows whether the patient is sick or not.
7. pregnant - Shows whether the patient is pregnant or not.
8. thyroid_surgery - Shows whether the patient has gone through Thyroid surgery or not
9. I131_treatment - Shows whether the patient is going through Radioactive Iodine(I-131) treatment or not.
10. query_hypothyroid - Shows whether the patient has raised any query regarding hypothyroid or not. Hypothyroid means thyroid is less than the optimum amounts.
11. query_hyperthyroid - Shows whether the patient has raised any query regarding hyperthyroid or not. Hyperthyroid means thyroid is more than the optimum amounts.
12. lithium - Shows whether the patient is on Lithium treatment or not.
13. goitre - Shows whether the patient is suffering from Goitre or not.
14. tumor - Shows whether the patient has any tumor or not. Tumor can either be benign or malignant.
15. hypopituitary - Shows whether the patient is on hypopituitary gland related treatment or not. 16. psych - Shows the patient’s psych evaluation.
17. TSH_measured - Shows whether the thyroid stimulating hormone(TSH) has been measured or not.
18. TSH - If TSH has been measured then this columm shows the value of the TSH.
19. T3_measured - T3 is one of two major hormones made by your thyroid. Shows whether T3 of the patient has been measured or not.
20. T3 - If T3 has been measured then this columm shows the value of the T3.
21. TT4_measured - T4 is another one of two major hormones made by your thyroid. Shows whether T4 of the patient has been measured or not.
22. TT4 - If T3 has been measured then this columm shows the value of the T4.
23. T4U_measured - Shows whether T4U has been measured or not.
24. T4U - If measured, the value of the same.
25. FTI_measured - Shows Whether FTI(Free Thyroxine Index) for the patient has been measured or not.
26. FTI - If FTI has been measured then this columm shows the value of the FTI.
27. TBG_measured - Shows whether Thyroxine-Binding Globulin Deficiency has been measured or not.
28. TBG - If TBG is measured then it shows value of the same.
29. referral_source - Shows the organization/hospital from where the patient has been refferred.
30. Class - This is the Dependent feature. It consists of 4 classes ‘negative’, ‘compensated hypothoroid’, ‘primary hypothoroid’, ‘secondary hypothoroid’.

This is a POC(Proof of concept) kind-of project. The data used here comes up with no guarantee from the creator. So, please don’t use it for detecting Thyroid class type for a patient. If you do so, the creator is not responsible for anything. However, this project presents the idea that how we can use ML into practice for healthcare sector.

MOTIVATION 💪

Healthcare domain has a lot of issue and Thyroid is one of the glands which gets impacted by diseases and makes a large part of the polulation go for testing, medication and surgery for the treatment of Thyroid related disorders. This is a small and humble project which is using multiple datasets and merges those to create a model and predict the type of Thyroid condition the patient has namely; “negative”, “compensated hypothyroid”, “primary hypothyroid”, “secondary hypothyroid”. In this project, I present a website in which the predictions are implemented as; Batch File PRediction(which predicts and generates a new file with predicted values based in the pre-trained model. For the visibility predicting application, the user can input the data and the application will predict the class of Thyroid out of the 4 classes - “negative”, “compensated hypothyroid”, “primary hypothyroid”, “secondary hypothyroid”.

Built with 🛠️

DEPLOYMENT 🚀

Deployment is done using deploy branch.
This website is deployed at AWS.
Github page: https://sagar61205.github.io/Thyroid-detection/
How to use?
Value-based-prediction ==> enter the corresponding values and it will fetch the number of calories burnt.
File-based-prediction ==> Click in ‘default file prediction’ to see the prediction on the already trained model OR Enter an absolute file path and clixk on ‘Custom file predict’.’.

Application screenshot for the ‘About’ section.

Application screenshot showing types of Thyroid disorders:

Application screenshot for batch file prediction. Here deafult prediction takes place on pre-trained dataset. A custom file file prediction can also be done.

Application screenshot for the important information section:

Prediction(Batch file):