MLCourse assignments

• Assignment 1
  • Exploratory Data Analysis (EDA) of Olympic games with Pandas Bonus
    Ten questions about 120 years of Olympic history:
    • Group By and Aggregation
    • Slicing
    • Dropping duplicates
    • Sorting
    • Feature Engineering
• Assignment 2
  • Exploratory Data Analysis (EDA) of US flights with Pandas, Matplotlib, and Seaborn Bonus
    The dataset has information about carriers and flights between US airports during the year 2008:
    • Group By and Aggregation
    • Slicing
    • Sorting
    • Counting
    • Mapping
    • Cross tabulation
    • Melting
    • Plotting
      • Hist
      • Plot
      • Rolling window
      • Box plot
      • Bar plot
• Assignment 3
  • Understanding Decision Trees for regression Bonus
    • Intuition for building decision trees for regression problems - step by step
    • Introduction (and implementation) of criterion for maximization, impurity of the node, calculation of leaf value
    • Prediction heart diseases using decision tree (feature engineering, data splitting, validation, hyperparameter search, feature importance)
  • Implementation of the decision tree algorithm Bonus
    • Implementation of the decision tree algorithm from scratch in scikit-learn style for classification and regression problems:
      • Leaf values calculation (mean, median, voting)
      • Criterion implementation (entropy, gini, variance, mad_median)
      • Building tree, predicting target, predicting probability
      • Testing implementation on real tasks (hyperparameter search, etc.)
• Assignment 7
  • Dimensionality reduction and clustering methods
    • Working with Samsung Human Activity Recognition dataset (6 clusters).
    • Scaling data
    • Applying PCA
    • Clustering with K-Means (choosing the best number of clusters)
    • Agglomerative clustering
    • Calculating the Adjusted Rand Index for two clustering algorithms
    • Solving this task using SVM with linear kernel
  • Principal Component Analysis and Clustering Bonus
• Assignment 8
  • Stochastic gradient descent for classification and regression Bonus
    • Implementation of linear regression (SGD) from scratch in scikit-learn style and testing it on height/weight dataset
    • Comparing from-scratch regression implementation with scikit-learn implementation
    • Implementation of logistic regression (SGD) with $L_2$ regularization from scratch in scikit-learn style and testing it on breast cancer UCI dataset
    • Solving movie review classification task (IMDB) using logistic regression and its SGD variation
      • Feature engineering using CountVectorizer
      • Understanding advantages of SGDClassifier