Prajwalamin / pca-model-comparison

Comparison of RandomForestClassifier and SGDClassifier when applied PCA performed on the MNIST dataset

PCA Model Comparison

Comparing the performance of Random Forest Classifier and SGD Classifier before and after applying dimensionality reduction algorithm (PCA) on the MNIST dataset.

Notebook

Results

• Random Forest Classifier:

  • Training Time (seconds):
    • Before PCA: 55
    • After PCA: 131
  • Accuracy:
    • Before PCA: 97%
    • After PCA: 95%

• SGD Classifier:

  • Training Time (seconds):
    • Before PCA: 175
    • After PCA: 38
  • Accuracy:
    • Before PCA: 90
    • After PCA: 88