/SBME_ML_PneumoniaClassification

Machine Learning Course - Final Project

Primary LanguageJupyter Notebook

SBME_MachineLearning_Project

Machine Learning Course - Final Project

Artificial Intelligence in Medicine

Pneumonia Classification Based on Chest X-rays

Firstly we load the data from their respective folder paths and store them in 2D arrays:
→ index[0] corresponds to the image array. → index[1] corresponds to the label of said image. → label = 0 corresponds to the "Pneumonia" class. → label = 1 corresponds to the "Normal" class.

Secondly we make a simple plot to take a look at its distribution
data distribution
We can notice that the data is fairly biased to Pneumonia. This means a few things:

  • Results will inevitably be biased to Pneumonia.
  • Overfitting will eventually occur.
  • Data manipulation is very needed for this data.

This can be solved in few ways:

  • Cutting the Pneumonia data to be equal to the Normal data.

    This isn't a viable option since the data quantity is already small.

  • Performing data augmentation on the Normal data to be equal to the Pneumonia data.

    Logically this seems like a good option, but may result in some overfitting to the available Normal features.

Since our data is of type image, the data augmentation methods need to fit the image criteria: data augmentation methods

Data augmentation libraries

  1. Augmentor → time consuming + high cpu usage
  2. Albumentations → used
  3. Imgaug → high cpu usage
  4. AutoAugment (DeepAugment) → errors in importing dependencies

Choice of number of PCA components

pca components
According to the cumulative sum plot of the obtained PCA components, we can see that the variance is almost a constant straight line after roughly 1500 components.
We picked PCA(n_components = 1000)

PCA components = 700, Image dimensions = (80,80):

Cutting the Pneumonia data to be equal to the Normal data.

Scores were as follows:

  • train score = 98.4%
  • test score = 80.0%

PCA components = 1000, Image dimensions = (100,100):

Augmenting the Normal data to be equal to the Pneumonia data.

Scores were as follows:

  • train score = 98.4%
  • test score = 78.0%

  1. svm classifier:

    • train score = 98.11%
    • test score = 78.37%

  2. logistic regression:

    • train score = 97.10%
    • test score = 74.68%

  3. KNN classifier:

    • train score = 96.01%
    • test score = 76.60%

  4. ensemble learning 1.0:

    • train score = 99.25%
    • test score = 77.56%

  5. ensemble learning 2.0 with gradient boosting:

    • train score = 98.66%
    • test score = 75.80%

Important Note

Basic machine learning algorithms in this field are considered outdated given the huge advancements in the field of artificial intelligence. Deep learning models should prove more helpful in such cases.