Data mining and machine learning mini-projects on UCI machine learning repository datasets
- Iris dataset EDA & visualization. Preprocessing the Iris dataset for tasks like classification and clustering, and visualizing it using different plots with seaborn and matplotlib.
- Decision tree, ensemble, & Naive Bayes classification on haberman dataset. Comparing different classification methods like DT, Naive Bayes, and also several ensemble techniques like Bagging, Boosting, And Adaboost on Haberman's Survival dataset to predict breast cancer patients' survival rate.
- Clustering on mall customer segmentation dataset. Clustering the mall customer segmentation dataset (Not from UCI. From Udemy's MachineLearningAZ course) using different clustering algorithms like K-Means and Agglomerative Hierarchical clustering.
- SVM classification on Breast cancer coimbra dataset. Comparing SVM with Polynomial, Sigmoid,and Gaussian kernels on breast cancer Coimbra dataset for classification and cancer detection.
- decision Tree & KNN classification on hayes roth & haberman datasets. Implementing DT classifier (and tuning with GridSearch) on Hayes Roth dataset, and KNN classifier with euclidean distance (also tuned with GridSearch) on Haberman's dataset while experimenting different aspects.
- Handwritten digits classification on semeion dataset using FCNN. Implementing Fully-connected neural network from scratch for in-depth understanding and using it to classify semeion hand-written digits, then comparing the performance with the Keras version (in a few lines of code).
- Handwritten digits classification on MNIST dataset using DCNN Classifying MNIST digits with CNN using Keras, and hyper-parameter tuning with a customized GUI bulit with tkinter (which takes hyper-parameter values from the user and tunes the CNN accordingly).