1. Home
  2. Sample Projects
  3. Medical diagnosis

Medical diagnosis

Print Friendly, PDF & Email

◷ Reading Time: 4 minutes

This example takes a dataset of symptoms to train the model and diagnose whether the patient has cold, flu or allergies. It uses the Naive Bayes algorithm for predictions.

The flow has three steps:

  • Input data
  • Train
  • Predict

In this example, we input the sample observed data as a CSV file. It contains the temperature, Humidity and the possibility of playing Golf. Using this dataset, the model will be created and trained to make predictions.

Open sample project

The “PredictiveAnalytics” project can be found on FlexRule Welcome screen once you open FlexRule Designer. If not, all of the sample projects can be found at the default sample project location:

C:\Users\<user name>\Documents\Pliant Framework\FlexRule Designer\<designer version>\Samples\PredictiveAnalytics

Then open Patients-NaiveBayes.xml

Run the Prediction Sample

1. Once you open Patients-NaiveBayes.xml, click on Run to run the project

2. Click OK on the Data Feed Provider window (There is no need to change options in this example)

3. The prediction will be shown in the Parameters window

Flow Description

1. Once you open Patients-NaiveBayes.xml, click on Debug to start running the project

2. Click on Next Step to go through the flow step by step.

3. Once it reaches the DataSource node,

you can see the properties; data source type, data file path, etc.

4. Click Next Step again and when you reach the train node,

5. Click on data in the Parameters window

and then click on Data viewer on the right side of Parameters window

and you will see the input data of the model.

6. Click Next Step and when you reach the predict node,

you can see the properties

7. Click on Prediction Input in the properties

to see the set of data we want to get the prediction for.

8. Click Next Step to end the flow and you will see the prediction as to the percentages of possible illness in the Parameters window.

Updated on June 24, 2019

Was this article helpful?

Related Articles