Perfectly Awesome

Data Mining - a Programmers Guide

• Data Mining: a Programmer’s Guide (Zacharski)

  • Intro

  • Recommendation Systems

    Intro; finding similar items; Manhattan distance; Euclidean distance; Minkowski distance; Pearson correlation coefficient; cosine similarity; k-nearest-neighbors in Python; book crossing dataset

  • Item-based Filtering

    Explicit & implicit ratings; user-based filters; item-based filters; adjusted cosine similarity; slope one algorithm; Python code; MovieLens dataset

  • Classification

    Pandora-like systems; selecting appropriate attributes; example; data normalization; modified standard score; Python code; sports example; acquiring attribute data

  • Classification, Pt2

    Training sets & test data; 10-fold cross validation; adding data vs algorithm tweaks; kNN; Python code

  • Naive Bayes & Probability Density Functions

    Lazy & eager learning; probability refresher; conditional probability; Bayes theorem; Python code; Congress Voting dataset; Gaussian distribution; Python code

  • Naive Bayes & unstructured text

    Positive & negative texts; classifier training; stop words; newsgroup classifier; Python code; sentiment analysis

  • Clustering

    Intro; hierarchical; single/complete/average linkages; dog breed clusters; breakfast cereal clusters; Kmeans; Kmeans++, Enron email dataset