SML-Assignments

Reports of all assignment are available in ./Report directory

Note:- Everything in assignments is implemented from scratch except some of the classifiers.

Assignment-1

• Classification on FMNIST Dataset.
• Performed Analysis of Naive Bayes Classifier.
• Performed Analysis for Classification on MNIST dataset.
• Performed K-Fold cross validation with analysis.
• Plotted various evaluation graphs like ROC, DET, Confusion Metric.

Assignment-2

• Performed Decorrelation on given dataset.
• Incorporated Risk Matrix for Naive Bayes classifier.
• Plotted Decision Boundary using mesh grid.
• Performed Classification with some missing data points value.

Assignment-3

• Performed Classification on Yale Face Dataset and CIFAR-10 Dataset.
• Implemented PCA & LDA from scratch.
• Did analysis using 5-Fold Cross Validation
• Analysis for PCA with varying Eigen Energy.
• Ensemble Learning:
  • Implemented Bagging from scratch using decision trees.
  • Implemented Ada-Boosting from scratch using decision tree classifier.
  • Performed Analysis with 5-Fold cross validation.
  • Analysis of using different Normalization techniqies like Z-Score, MinMax Normalization and Tanh.

Assignment-4

• Implemented generic layers Neural Network from scratch.
• Analysis with different Activation functions.
• Implemented Auto Encoder using PyTorch on MNIST Dataset. Trained a neural network using the reduced dimension from Auto Encoder.
• Hands on with Panda Library.
• Did Feature Selection, Extraction and Feature Clustering for labeled data and unlabeled data.

Bonus Assignment

• Private Kaggle Competition of Image Classification.
• Data included 20 classes and 10K Train Samples with 1K Test samples.