ML assignment using a DIGITS dataset from sklearn. Initially we perform some dataset analysis tasks, followed by principal component analysis using sklearn and the implementation of k-nn sklearn algorithm. We present how the algorithms performance changes by adding noise to the dataset.
Implementation of Logistic Regression for the IRIS sklearn dataset, performance analysis of the algorithms, use of PCA for dimensionality reduction, classification report and confusion matrix.
Principal Component Analysis. The notebook contains: i) Implementation of PCA to the digits dataset from scikit-learn. ii) Noise and Dimensionality Reduction on the dataset. iii) Eigenfaces of the dataset.
Lasso and Ridge implementation. Coefficients optimization and error analysis.
Implementation of knn adn nested cross validation using low level code. We also present a report discussing our findings.