https://github.com/asadiahmad/image-classification-lda-and-pca

Image Classification with Perceptron and LDA and PCA dimension reduction
https://github.com/asadiahmad/image-classification-lda-and-pca

dimension-reduction image image-classification linear-discriminant-analysis machine-learning machine-learning-algorithms mnist mnist-classification perseptron perseptron-classification principal-component-analysis

Last synced: 7 months ago
JSON representation

Image Classification with Perceptron and LDA and PCA dimension reduction

Host: GitHub
URL: https://github.com/asadiahmad/image-classification-lda-and-pca
Owner: AsadiAhmad
License: mit
Created: 2024-11-12T21:10:55.000Z (11 months ago)
Default Branch: main
Last Pushed: 2024-11-12T21:12:40.000Z (11 months ago)
Last Synced: 2024-11-12T22:21:34.667Z (11 months ago)
Topics: dimension-reduction, image, image-classification, linear-discriminant-analysis, machine-learning, machine-learning-algorithms, mnist, mnist-classification, perseptron, perseptron-classification, principal-component-analysis
Language: Jupyter Notebook
Homepage:
Size: 274 KB
Stars: 0
Watchers: 1
Forks: 0
Open Issues: 0
Metadata Files:
- Readme: README.md
- License: LICENSE

Awesome Lists containing this project

README

# Image-Classification-LDA-and-PCA
Image Classification with Perceptron and LDA and PCA dimension reduction

## Tech :hammer_and_wrench: Languages and Tools :

## Run the Notebook on Google Colab

You can easily run this code on google colab by just clicking this badge [![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/AsadiAhmad/Image-Classification-LDA-and-PCA/blob/main/Image_Classification_with_LDA_%26_PCA.ipynb)

## Conclusion

In this project, we demonstrated the application of LDA and PCA for effective dimensionality reduction in image classification tasks. In out analysis we found the LDA better than the PCA for dimension reduction task. Accuracy of the LDA is much better than PCA because LDA is supervised Learning and PCA is unsupervised learning method. Most of the time supervised Learning have better accuracy than unsupervised learning methods because supervised method using the labels but unsupervised does not.However when we don't have any labels then we can just use the unsupervised method.

Here you can see BoxPlot of each method we have been used :

### LDA BoxPlot

### PCA BoxPlot

You can see better accuracy at LDA than the PCA method. Another conclusion we can get is look like when dimension reach a point like 9 to 25 dimension in PCA method our accuracy does not grow much show that our method is reaching its highest point of accuracy after that maybe we have overfit and you can see the overfitting and underfitting results down here :

### LDA Results Table