Group project for Machine Learning and Deep Learning ECE642 subject.
Task - Apply PCA on Optdigits dataset. Choose features that explain 90 % of the variance. Build a classifier using K-means, and calculate accuracy.
PLS NOTE THAT USING ANY INBUILT LIBRARIES EXCEPT NUMPY is not allowed.
- Explanation of problem and dataset: 5 Marks
- Code: 5 Marks
- Proper Coding style - 2 marks
- Suitable Variable Names - 1 mark
- Comments - 1 mark
- Indentation - 1 mark
- Results and Discussion with images, graphs etc : 10 Marks
The file named as projectnotebook.ipynb
has the entire code.
Team Members -
- Rishabh Bhatt (1MS18EC089)
- Yash Bagri (1MS18EC134)
- Khushi S Patil (1MS18EC046)
Link to dataset - https://archive.ics.uci.edu/ml/datasets/optical+recognition+of+handwritten+digits or https://datahub.io/machine-learning/optdigits
Results - Successfully applied PCA to the original dataset and then applied the Kmean clustering algorithm to obtain an accuaracy of around 70%.
Link to notebook - https://github.com/WrathTitan/OptdigitsPCA-Kmeans/blob/master/projectnotebook.ipynb
Note: You can run the project on Google Colab or in your system with python3 installed along with the following dependencies
run in terminal to install the following - pip3 install <Name-of-the-library>
numpy
pandas
sklearn
matplotlib
seaborn