Application of Boosted Tree Classifier for Predicting Disease from Symptoms

Research based testing Boosted Tree Classifier for Predicting Disease from Symptoms

Dependencies

1. python 64 bit (Strictly) (32 bit will raise errors) If you have 32 bit version. Uninstall it first then install 64 bit version.
2. tensorflow 2.4
3. pandas

Note: Ignore 10000 Warnings. Warnings are because we are using a version of tensorflow which can run without GPU. If you have GPU update cuda toolkit and cuDNN to run model faster.

If you encounter an error named Value out of range that is because you put a value of variable that is way too large.

Variables that needs to be tweaked

Edit these variables in main.py file only

• noOfTree
• maxTreeDepth
• learningRate
• noOfBatchesPerLayer

Note your observation in the observation.xlsx file under the sheet of YOUR NAME

This algorithm uses plenty of randomness to create first trees so everytime you run the model even with same variable values you might get different values.

Videos I referred to understand Boosted Tree Classifier

In Sequence

1. https://youtu.be/3CC4N4z3GJc
2. https://youtu.be/2xudPOBz-vs
3. https://youtu.be/jxuNLH5dXCs
4. https://youtu.be/StWY5QWMXCw

You can choose not to watch these videos and find your own tutorials.

Reading content

• Tensorflow Documentation: https://www.tensorflow.org/api_docs/python/tf/estimator/BoostedTreesClassifier
• Example of Boosted Tree: https://medium.com/tensorflow/how-to-train-boosted-trees-models-in-tensorflow-ca8466a53127

Research Paper Reference

• https://examples.yourdictionary.com/basic-research-paper-format-examples.html