By the end of the course, you should know the following key concepts:

Uses of Data Mining
    examples, industry areas
    CRM, targeted marketing, personalization
    strategic use (e.g., Capital One case)
    roles of data mining vs. experimental design

Data Warehousing
  relational vs. transactional (OLAP vs. OLTP)
  OLAP functions: 
    slice, dice, roll-up, drill down
  star and snowflake schemas
  ETL (Extract Transform Load) 
      meta-data, importance of
  problems of data quality

Methods
  visualization (PTDD)
    e.g., mosaic plots, interactive visualization (e.g., Spotfire)
  clustering
     kmeans, k-nearest neighbors, agglomerative clustering
  decision trees
    splitting, pruning
  linear and logistic regression
    independence, colinearity, interactions
    stepwise regression
  artificial neural networks
  support vector machines (SVM)
  categorical data
     how to use in regression methods
  missing data
    missing at random / not at random
    how to handle in regression, decision trees

Evaluation
  accuracy and its limitations
  precision/ recall, lift, confusion matrix
  overfitting
     cross-validation, test/train, in-sample/hold-out
  causality vs. correlation
     lurking variables and the role of time
     (e.g. gazelle.com case)
  evaluation of model is not business evaluation
     "type 3 errors"

Process
  CRISP methodology
    Business Understanding
    Data Understanding
    Data Preparation
    Data cleansing
    Evaluation
    Deployment
  Common causes for Data Mining project failure
    Failure to define the business problem
    Failure to get key people bought into the project early
    Insufficient Data size or quality
      - insufficient effective number of observations
      - missing key features
   Tools and their attributes

Text Mining and Search
  information retrieval vs. information extraction
  collaborative filtering
  pagerank
  named entity recognition
  why text mining is hard