Thyroid Classification

This application builds a multi-class classification methodology to predict the type of Hypothyroidism based on the given training data. Refer to the application workflow below for details. The code for this app can be found here.

image

Image source

Dataset details:

The dataset contains 30 columns with features such as age, sex, TSH, T3, T4 etc. The target variable contains the following three classes: "negative, compensated hypothyroid, primary hypothyroid". Datasets for training and prediction are downloaded from an S3 bucket and can be found in the Raw_Data folder. A 'schema' file is used to verify the data format. Logging is performed frequently and logs are stored in the specific log files.

Exploratory Data Analysis performed on the data can be found in the notebook file here.


Build Models and Generate Predictions:

Generate Predictions:

Various machine learning models (KNN, SVM, Random Forest, AdaBoost and XGBoost) have already been built using training data. Click the button to generate predictions instantly.


The table below shows a subset of the data used for generating predictions.

{% for table in tables %} {{ table|safe }} {% endfor %}

Train the Models:



Application Details:

Libraries Used:

image